Skip to main content
Sign in Create a free account
Varsity Tutors
Sign in
Back to lessons

Managerial Accounting • Managerial Accounting Foundations

High-Low Method

A quick estimation technique that separates mixed costs into their fixed and variable components using only two data points.

Explore the lesson 16 min read
Managerial Accounting emblem with a calculator, a financial report, and a column chart.
Inside this lesson 10 sections 5 questions 9 figures

SECTION 1

Historical Context & Motivation

Understanding how costs behave in response to changes in business activity is one of the oldest and most essential challenges in managerial accounting. As organizations grew in scale during the industrial era, managers needed reliable ways to forecast costs for budgeting, pricing, and performance evaluation. The fundamental problem was that most business costs are mixed costs—they contain both a fixed component that remains constant regardless of activity and a variable component that fluctuates proportionally with output. Separating these two elements became a critical prerequisite for managerial decision-making.

The High-Low Method emerged as one of the earliest practical techniques for estimating cost behavior. While statisticians had long used regression analysis to fit lines to data, such methods required computational resources that were unavailable to most practicing accountants until the late twentieth century. The High-Low Method offered an elegant shortcut: by selecting only the highest and lowest activity levels from a dataset, a manager could derive a linear cost function with nothing more than pencil and paper. This approach became a staple of cost accounting textbooks and remains widely taught as both a standalone estimation tool and a conceptual stepping-stone toward more sophisticated methods such as least-squares regression.

1900s–1920s • Industrial Cost Accounting Emerges

Mass production in factories drives demand for systematic cost classification. Accountants begin distinguishing between fixed, variable, and semi-variable (mixed) costs to improve budgeting accuracy.

1930s–1950s • Cost-Volume-Profit Analysis Takes Shape

The break-even chart and CVP framework become standard managerial tools. Practitioners adopt the High-Low Method as a quick, manual technique for estimating the variable rate and fixed cost within mixed-cost accounts.

1960s–1980s • Textbook Canonization

Leading cost accounting textbooks by Horngren, Garrison, and others codify the High-Low Method alongside scattergraph analysis and regression, establishing the trio of cost estimation techniques still taught today.

1990s–Present • Spreadsheet Era and Comparative Context

With spreadsheet software making regression analysis trivial, the High-Low Method shifts in pedagogical role—valued less for daily use and more for building intuition about cost behavior and illustrating the limitations of simplified estimation.

The central question the High-Low Method addresses is straightforward yet vital: given a set of historical observations of total cost at various activity levels, how can a manager quickly estimate the variable cost per unit of activity and the total fixed cost? The method's answer is to let the two most extreme data points define the cost line—an approach that is remarkably easy to execute but carries important assumptions and limitations that every business student should understand.

SECTION 2

Core Principles & Definitions

Before applying the High-Low Method, it is important to internalize several foundational ideas about cost behavior. Managerial accountants classify costs by their responsiveness to changes in an activity driver—the measurable factor (such as machine hours, units produced, or miles driven) that causally influences the cost. A variable cost changes in total proportionally with the activity driver, while a fixed cost remains constant in total across the relevant range. A mixed cost (also called a semi-variable cost) contains elements of both, and the High-Low Method's purpose is to decompose it into those two components.

1

Linear Cost Assumption

The method assumes that total cost behaves linearly within the relevant range—meaning the relationship between activity and cost can be represented by a straight line: Y = a + bX.
2

Two-Point Estimation

Only the data points at the highest and lowest activity levels are used. Because two points uniquely determine a line, no additional observations influence the estimate.
3

Activity-Based Selection

High and low points are identified by the activity level (the independent variable X), not by the total cost (Y). This distinction prevents selecting outlier cost observations that do not reflect normal operations.
4

Slope = Variable Cost Rate

The slope of the line connecting the two extreme points equals the variable cost per unit of activity (b). Once known, the fixed cost (a) is found by substitution into either data point.
5

Relevant Range Caveat

The resulting cost equation is valid only within the span of observed activity levels. Extrapolating beyond this range ignores potential step costs, capacity constraints, and non-linear behavior.
✦

KEY TAKEAWAY

Think of the High-Low Method like estimating the slope of a hiking trail by measuring only the elevation at the trailhead and the summit. You get a single average gradient that ignores every dip and rise in between—useful for a rough gauge of difficulty, but potentially misleading if a steep cliff occupies the middle section. Similarly, the High-Low Method gives you a quick average variable cost rate, but it cannot detect non-linearities or unusual data patterns hidden between the extremes.

SECTION 3

Visual Explanation

The best way to grasp the High-Low Method is to visualize it on a scattergraph—a plot where the horizontal axis represents the activity level and the vertical axis represents total cost. Each historical observation is plotted as a point, and the High-Low Method draws a straight line through the two extreme activity points, ignoring every other observation. The diagram below illustrates this process with a hypothetical dataset of monthly machine hours and maintenance costs.

High-Low Method on a ScattergraphActivity Level (Machine Hours)Total Cost ($)2004006008001,0001,2002,0004,0006,0008,00010,00012,000LOWHIGHHigh-Low LineY = a + bXLow: (300, 3,000)High: (1,100, 8,500)
The scattergraph plots nine months of data. The low point (300 hours, $3,000) and the high point (1,100 hours, $8,500) define the dashed cost line. All other observations are ignored by the High-Low Method.

Notice how the dashed line passes through only the two circled points. The remaining seven observations may fall above or below the line, suggesting that the High-Low estimate is an approximation rather than an optimized fit. In contrast, a least-squares regression line would minimize the total squared deviations across all nine points, generally yielding a more reliable estimate. Nonetheless, the visual simplicity of the High-Low Method makes it an excellent teaching tool for understanding the concept of slope as variable cost per unit and the y-intercept as fixed cost.

SECTION 4

Mathematical Framework

The High-Low Method rests on the standard linear cost equation. By treating cost as a dependent variable (Y) and activity as an independent variable (X), we express total mixed cost as a linear function. The method solves for the two unknowns—variable cost per unit (b) and total fixed cost (a)—in two sequential steps.

MIXED COST EQUATION

Y=a+bXY = a + bX

Where Y = total mixed cost, a = total fixed cost (y-intercept), b = variable cost per unit of activity (slope), X = activity level.

STEP 1 — VARIABLE COST PER UNIT (SLOPE)

b=(Yhigh−Ylow)/(Xhigh−Xlow)b = (Y_high − Y_low) / (X_high − X_low)

The change in total cost divided by the change in activity level yields the variable cost rate. This is the classic rise-over-run slope formula applied to the two extreme data points.

STEP 2 — TOTAL FIXED COST (Y-INTERCEPT)

a=Yhigh−b×Xhigh(orequivalently,a=Ylow−b×Xlow)a = Y_high − b × X_high (or equivalently, a = Y_low − b × X_low)

After computing b, substitute into either the high or low data point and solve for a. Both substitutions must yield the same fixed cost.

STEP 3 — COST PREDICTION

Yestimated=a+b×XpredictedY_estimated = a + b × X_predicted

Once a and b are known, the cost equation can be used to predict total cost for any activity level within the relevant range. This is the practical payoff: budgeting, variance analysis, and break-even calculation.

⚠

Common Pitfall

A frequent student error is selecting high and low points based on the cost column rather than the activity column. The High-Low Method defines 'high' and 'low' strictly by the independent variable (activity level). If the month with the highest cost does not coincide with the highest activity, the cost figure at the highest activity is still the correct choice.

SECTION 5

Step-by-Step Process Flow

The following diagram presents the complete High-Low Method process as a flowchart. It begins with data collection and moves through point selection, slope calculation, intercept determination, and finally cost prediction. Refer to this visual as a procedural checklist whenever you apply the method.

High-Low Method — Process Flowchart1. Collect Historical Data2. Identify High & Low Activity PointsSelect by activity level (X), NOT by cost (Y)3. Calculate Variable Cost Rate (b)b = (Y_high − Y_low) ÷ (X_high − X_low)4. Solve for Fixed Cost (a)a = Y_high − b × X_high5. Construct Cost EquationY = a + bX → Predict costs
The five-step flowchart moves from raw data collection to a usable cost equation. The critical decision—selecting points by activity level, not cost—is emphasized in Step 2.

The process assumes that the dataset has already been screened for outliers—abnormal observations caused by one-time events such as equipment breakdowns, strikes, or natural disasters. If the highest or lowest activity period happens to be an outlier, the resulting cost equation will be distorted. Best practice is to examine a scattergraph before applying the method, visually confirming that the extreme points are consistent with the overall data pattern. If an outlier is detected at an extreme, it should be excluded and the next most extreme observation used instead.

SECTION 6

Worked Example

Greentree Manufacturing wants to estimate the fixed and variable components of its monthly maintenance cost. The company's controller has compiled the following six-month dataset relating machine hours (the cost driver) to total maintenance cost.

MonthMachine Hours (X)Total Maintenance Cost (Y)
January1,500$10,000
February2,500$14,000
March2,200$12,500
April3,000$16,500
May1,800$11,000
June2,800$15,500

Greentree Manufacturing — Six-Month Maintenance Cost Data

Applying the High-Low Method to Greentree Manufacturing

  1. 1

    Step 1 — Identify the High and Low Activity Points

    Scan the machine hours column (X) to find the highest and lowest activity levels. The highest activity is 3,000 hours in April with a cost of $16,500. The lowest activity is 1,500 hours in January with a cost of $10,000.
    High Point: (3,000, $16,500) | Low Point: (1,500, $10,000)
  2. 2

    Step 2 — Calculate the Variable Cost per Machine Hour (b)

    Apply the slope formula: b = (Y_high − Y_low) ÷ (X_high − X_low) = ($16,500 − $10,000) ÷ (3,000 − 1,500) = $6,500 ÷ 1,500.
    b = $4.333 per machine hour (≈ $4.33)
  3. 3

    Step 3 — Calculate the Total Fixed Cost (a)

    Substitute into the high point: a = Y_high − b × X_high = $16,500 − ($4.333 × 3,000) = $16,500 − $13,000 = $3,500. Verification with the low point: a = $10,000 − ($4.333 × 1,500) = $10,000 − $6,500 = $3,500. Both yield the same value, confirming the computation.
    a = $3,500 per month (fixed cost)
  4. 4

    Step 4 — Construct the Cost Equation

    Assemble the mixed cost equation: Y = $3,500 + $4.333X. This equation can now be used for prediction.
    Total Maintenance Cost = $3,500 + $4.333 × Machine Hours
  5. 5

    Step 5 — Predict Cost for a New Activity Level

    If Greentree plans to operate at 2,600 machine hours next month, the estimated maintenance cost is: Y = $3,500 + $4.333 × 2,600 = $3,500 + $11,267 = $14,767.
    Estimated cost at 2,600 hours ≈ $14,767

SECTION 7

Strengths & Limitations

No estimation technique is universally superior; each involves trade-offs between simplicity, accuracy, and data requirements. Understanding the strengths and limitations of the High-Low Method helps managers judge when to deploy it and when to invest in more rigorous alternatives.

DimensionStrengthsLimitations
SimplicityRequires only basic arithmetic—no software, no statistical training. Can be performed in minutes with a calculator.Oversimplification ignores the richness of the full dataset. A single outlier at an extreme can skew results dramatically.
Data RequirementNeeds only two observations—the minimum possible for defining a line. Useful when data is scarce.Ignores all intermediate data points, which may contain important information about the true cost behavior pattern.
AccuracyProvides a reasonable first approximation when data points cluster along a clear linear trend.No goodness-of-fit metric (R²). The user has no way to assess how well the line fits without comparing to other methods.
SensitivityQuick screening tool—if the high-low estimate is wildly different from expectations, it signals potential data issues.Results change entirely if different extreme points are chosen (e.g., after removing an outlier), making the estimate unstable.
Pedagogical ValueBuilds conceptual understanding of mixed-cost decomposition before introducing regression analysis.May create a false sense of precision if students do not learn its limitations alongside the procedure.

Strengths vs. Limitations of the High-Low Method

✦

KEY TAKEAWAY

Think of the High-Low Method as the equivalent of estimating a city's average temperature by looking only at the hottest and coldest days of the year. You get a rough midpoint, but you miss the distribution of temperatures in between—some months might be consistently mild, others erratic. Similarly, the High-Low Method can give you a serviceable first estimate of cost behavior, but for high-stakes decisions like product pricing or capital budgeting, you should always supplement it with regression analysis or other statistical techniques that consider the full dataset.

SECTION 8

Connection to Regression Analysis & Advanced Cost Estimation

The High-Low Method is best understood as a gateway to more sophisticated cost estimation techniques. In practice, managerial accountants frequently use least-squares regression (also called the method of ordinary least squares, or OLS) to estimate the same linear cost equation. Regression uses all available data points and minimizes the sum of squared residuals, producing statistically optimal estimates of the slope and intercept. Understanding the High-Low Method first makes the motivation for regression intuitively clear: if two points can define a line, then many points can define a better line.

FeatureHigh-Low MethodLeast-Squares Regression
Data Points UsedOnly the two extreme activity observationsAll available observations
Objective FunctionNone—simply connects two pointsMinimizes the sum of squared residuals (Σe²)
Goodness-of-Fit MetricNot availableR² (coefficient of determination) and adjusted R²
Outlier SensitivityVery high—one extreme outlier changes the entire lineModerate—outliers are diluted across many observations
Multiple DriversCannot accommodate multiple cost driversMultiple regression handles several independent variables
ComputationManual—pencil and calculatorTypically software-assisted (Excel, statistical packages)

High-Low Method vs. Least-Squares Regression

Beyond regression, advanced managerial accounting courses explore account analysis (a qualitative approach where managers classify each cost account as fixed, variable, or mixed based on judgment) and engineering estimates (using time-and-motion studies or material specifications to build cost functions from the ground up). Activity-based costing (ABC) takes the concept further by identifying multiple cost pools and drivers, allowing for far more granular cost prediction. Each technique occupies a point on the trade-off frontier between simplicity and precision, and the High-Low Method sits squarely at the simplicity end—making it a natural starting point for any cost estimation discussion.

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

Explain why the High-Low Method selects data points based on the activity level (X) rather than the total cost (Y). What problem could arise if a manager selected points based on cost instead?
Show answer
The High-Low Method aims to measure how cost responds to changes in activity—that is, the slope of the cost line. The independent variable (X) must therefore drive the selection so that the widest range of the causal factor is captured. If points were selected by the highest and lowest cost, the analyst might pick an observation where cost was unusually high due to a one-time event (e.g., an equipment failure) that did not correspond to the highest activity level. This would produce a slope that reflects random cost variation rather than the underlying variable cost rate, yielding a misleading cost equation.

Question 2: BASIC CALCULATION

A delivery company recorded shipping costs of $8,200 at 400 deliveries (low) and $14,600 at 1,000 deliveries (high). Using the High-Low Method, determine the variable cost per delivery and the total fixed cost.
Show answer
Variable cost per delivery: b = ($14,600 − $8,200) ÷ (1,000 − 400) = $6,400 ÷ 600 = $10.667 per delivery. Fixed cost: a = $14,600 − ($10.667 × 1,000) = $14,600 − $10,667 = $3,933. The cost equation is Y = $3,933 + $10.667X. Verification with the low point: $3,933 + ($10.667 × 400) = $3,933 + $4,267 = $8,200, which matches the given data.

Question 3: INTERMEDIATE

Riverside Corp. collected eight months of data on labor hours and utility costs. The highest activity was 5,000 labor hours at $23,000 total utility cost, and the lowest was 2,000 labor hours at $14,000. Using the High-Low Method, estimate the utility cost for a month in which 3,800 labor hours are expected. Then explain why the estimate might be unreliable if one of the extreme months experienced a heat wave that dramatically increased cooling costs.
Show answer
Variable cost per labor hour: b = ($23,000 − $14,000) ÷ (5,000 − 2,000) = $9,000 ÷ 3,000 = $3.00 per labor hour. Fixed cost: a = $23,000 − ($3.00 × 5,000) = $23,000 − $15,000 = $8,000. Estimated utility cost at 3,800 hours: Y = $8,000 + $3.00 × 3,800 = $8,000 + $11,400 = $19,400. However, if the high-activity month coincided with a heat wave, the $23,000 figure includes abnormal cooling costs unrelated to labor hours. This inflates the slope (b) and distorts the fixed cost estimate. The correct approach would be to exclude the outlier and select the next-highest activity month as the high point.

Question 4: APPLIED

You manage a call center and have the following quarterly data: Q1: 12,000 calls handled, $95,000 total cost; Q2: 18,000 calls, $128,000; Q3: 22,000 calls, $152,000; Q4: 15,000 calls, $110,000. Using the High-Low Method, (a) determine the cost equation, (b) estimate total cost for a quarter with 20,000 calls, and (c) compute the expected total cost if the center scales up to 25,000 calls. Discuss whether the estimate in part (c) is reliable.
Show answer
(a) High point: Q3 (22,000 calls, $152,000); Low point: Q1 (12,000 calls, $95,000). Variable cost per call: b = ($152,000 − $95,000) ÷ (22,000 − 12,000) = $57,000 ÷ 10,000 = $5.70 per call. Fixed cost: a = $152,000 − ($5.70 × 22,000) = $152,000 − $125,400 = $26,600. Cost equation: Y = $26,600 + $5.70X. (b) At 20,000 calls: Y = $26,600 + $5.70 × 20,000 = $26,600 + $114,000 = $140,600. (c) At 25,000 calls: Y = $26,600 + $5.70 × 25,000 = $26,600 + $142,500 = $169,100. This estimate is unreliable because 25,000 calls exceeds the observed relevant range of 12,000–22,000 calls. At 25,000 calls, the center may require additional staff, equipment, or shifts, introducing step costs that the linear equation cannot capture.

Question 5: CRITICAL THINKING

Suppose two analysts apply the High-Low Method to the same 12-month dataset but obtain different cost equations. Analyst A uses months with the highest and lowest production volumes, while Analyst B excludes the highest-volume month as an outlier (a machine ran 24/7 due to a rush order) and substitutes the second-highest month. Critically evaluate whose equation is likely more representative of normal cost behavior, and propose a procedure that could objectively settle the disagreement.
Show answer
Analyst B's equation is likely more representative because the rush-order month reflects atypical operating conditions—overtime premiums, accelerated depreciation, and abnormal supply usage would inflate the cost at that volume, producing an artificially steep slope and an underestimated fixed cost. Analyst A's equation mechanically follows the High-Low procedure but violates the underlying assumption that observations reflect normal operations within the relevant range. To objectively resolve the dispute, both analysts should perform a least-squares regression on all 12 data points and examine the residuals. The rush-order month would likely show a large positive residual, confirming its outlier status. A regression excluding the outlier (11 data points) would provide a statistically grounded cost equation with an R² value, confidence intervals, and standard errors—tools that the High-Low Method simply cannot offer. This exercise illustrates the fundamental limitation of the High-Low Method: it provides no diagnostic information about data quality, whereas regression analysis allows the analyst to test assumptions and quantify uncertainty.

SUMMARY

Lesson Summary

Summary

The High-Low Method is a straightforward technique for decomposing a mixed cost into its fixed and variable components. It works by selecting the observations with the highest and lowest activity levels (not cost levels) and computing the slope (variable cost per unit) as the change in cost divided by the change in activity. The y-intercept (fixed cost) is then found by substitution. The resulting linear equation Y = a + bX enables cost prediction for any activity level within the relevant range.

While valued for its simplicity and speed, the method carries significant limitations: it ignores all data points between the extremes, provides no goodness-of-fit measure, and is highly sensitive to outliers at the extremes. For high-stakes decisions, it should be supplemented or replaced by least-squares regression, which uses all observations and provides statistical diagnostics. Nonetheless, mastering the High-Low Method builds essential intuition about cost behavior analysis and prepares students for more advanced techniques in managerial and cost accounting.

Varsity Tutors • Managerial Accounting • High-Low Method