Cost Behavior: High-Low Method & Regression Analysis

TAKEAWAYS:

• Cost behavior describes how costs change with production volume, classified as fixed, variable, or mixed.
• The High-Low Method separates mixed costs into fixed and variable components using the highest and lowest activity levels.
• Regression Analysis provides a more accurate cost estimation by fitting a line to all data points, minimizing errors.
• Cost equations (Y = a + bX) help predict total costs at different production levels, crucial for budgeting and pricing.
• Real-world applications include Daraz’s warehouse cost planning, Ncell’s network maintenance expenses, and Kathmandu’s traffic management costs.
• Exam focus: Expect numerical problems on segregating costs, predicting future costs, and interpreting regression outputs.

---

1. Understanding Cost Behavior

Cost behavior refers to how costs change in response to changes in production or sales volume. It helps businesses predict expenses and optimize operations. Costs are broadly classified into:

Types of Costs Based on Behavior

Why it matters:

• Fixed costs (e.g., rent for a Daraz warehouse) remain constant regardless of how many orders are processed.
• Variable costs (e.g., packaging for each Daraz order) increase with sales volume.
• Mixed costs (e.g., Ncell’s electricity bill) have both fixed and variable elements.

---

2. High-Low Method: Segregating Mixed Costs

The High-Low Method is a simple technique to separate mixed costs into fixed and variable components using the highest and lowest activity levels.

How It Works

1. Identify the highest and lowest activity levels (e.g., units produced or sales volume).
2. Calculate the variable cost per unit using the formula: $$\text{Variable Cost per Unit}=\frac{\text{Change in Cost}}{\text{Change in Activity}}$$
3. Determine the fixed cost by plugging the variable cost into the cost equation: $$Y=a+bX$$ where:
   • $Y$ = Total Cost
   • $a$ = Fixed Cost
   • $b$ = Variable Cost per Unit
   • $X$ = Activity Level

Worked Example: Kathmandu Retail Shop

Scenario: A retail shop in Kathmandu tracks its utility costs (electricity + water) for 6 months:

Step 1: Identify highest and lowest activity levels.

• Highest: 1,500 units (Aswin)
• Lowest: 500 units (Baisakh)

Step 2: Calculate variable cost per unit. $$\text{Change in Cost}=45,000-25,000=20,000\text{ Rs}$$ $$\text{Change in Activity}=1,500-500=1,000\text{ units}$$ $$\text{Variable Cost per Unit}=\frac{20,000}{1,000}=20\text{ Rs/unit}$$

Step 3: Calculate fixed cost using the highest activity level. $$Y=a+bX$$ $$45,000=a+20\times 1,500$$ $$45,000=a+30,000$$ $$a=15,000\text{ Rs (Fixed Cost)}$$

Cost Equation: $$Y=15,000+20X$$

Verification: For 800 units (Jestha): $$Y=15,000+20\times 800=15,000+16,000=31,000\text{ Rs (Close to actual 30,000 Rs)}$$

Limitations of High-Low Method:

• Ignores intermediate data points, leading to potential inaccuracies.
• Assumes a linear relationship, which may not always hold.

---

3. Regression Analysis: A More Accurate Approach

Regression Analysis fits a least-squares line to all data points, minimizing errors and providing a more precise cost equation.

Key Concepts

• Dependent Variable (Y): Total Cost
• Independent Variable (X): Activity Level (e.g., units produced)
• Regression Line: $Y=a+bX$, where:
  • $a$ = Intercept (Fixed Cost)
  • $b$ = Slope (Variable Cost per Unit)

Advantages Over High-Low Method

Worked Example: Ncell’s Network Maintenance Costs

Scenario: Ncell tracks its monthly network maintenance costs (Rs) against the number of active subscribers (in thousands):

Using Excel Regression:

1. Plot $X$ (Subscribers) vs. $Y$ (Cost).
2. Run linear regression to get:
   • Slope ($b$) = 1,500 Rs per 1,000 subscribers
   • Intercept ($a$) = 45,000 Rs (Fixed Cost)

Cost Equation: $$Y=45,000+1,500X$$

Prediction: For 75,000 subscribers ($X=75$): $$Y=45,000+1,500\times 75=45,000+112,500=157,500\text{ Rs}$$

Why Regression is Better:

• Accounts for all data points, reducing bias from outliers.
• Provides R² value (goodness-of-fit), e.g., R² = 0.98 means 98% of cost variation is explained by subscribers.

---

4. Real-World Applications

1. Daraz’s Warehouse Cost Planning

• Idea Used: Mixed Costs (High-Low Method)
• How: Daraz’s warehouse rent is fixed, but packing and shipping costs vary with orders.
  • Fixed Cost: Rent (Rs 500,000/month).
  • Variable Cost: Rs 20 per order.
  • Cost Equation: $Y=500,000+20X$.
  • Use: Helps Daraz set pricing and predict profits at different order volumes.

2. Ncell’s Network Expansion Decisions

• Idea Used: Regression Analysis
• How: Ncell uses regression to estimate maintenance costs for new towers.
  • Data: Past costs vs. subscriber growth.
  • Output: Predicts costs for 100,000 new subscribers.
  • Use: Justifies investment in new infrastructure.

3. Kathmandu Traffic Management (Unavoidable Cost)

• Idea Used: Opportunity Cost
• How: The Kathmandu Metropolitan City (KMC) faces traffic congestion.
  • Unavoidable Cost: Existing road maintenance (cannot be eliminated).
  • Opportunity Cost: Lost productivity due to delays (e.g., Rs 500 million/day).
  • Decision: KMC prioritizes flyovers over minor road repairs to maximize efficiency.

---

5. Comparing High-Low and Regression

---

6. Numerical Problems and Solutions

Problem 1: Segregating Costs (High-Low Method)

A manufacturing company provides:

Solution:

1. Highest: 7,000 units, Rs 60,000 Lowest: 3,000 units, Rs 40,000
2. Variable Cost per Unit: $$\frac{60,000-40,000}{7,000-3,000}=\frac{20,000}{4,000}=5\text{ Rs/unit}$$
3. Fixed Cost: $$Y=a+5X⟹60,000=a+5\times 7,000⟹a=25,000\text{ Rs}$$
4. Cost Equation: $$Y=25,000+5X$$

Problem 2: Predicting Costs (Regression)

Given regression output for a company:

• Intercept ($a$) = 10,000 Rs
• Slope ($b$) = 3 Rs/unit
• R² = 0.95

Questions: a) What is the fixed cost? Answer: Rs 10,000 (intercept). b) What is the variable cost for 5,000 units? $$Y=10,000+3\times 5,000=25,000\text{ Rs}$$ c) Interpret R² = 0.95. Answer: 95% of cost variation is explained by activity level; model is highly reliable.

---

7. Exam Tip

What Examiners Look For:

1. Correct Identification:

   • Clearly label fixed, variable, and mixed costs in problems.
   • Example: In the Kathmandu shop example, utility costs are mixed.

2. Step-by-Step Calculations:

   • Show all steps for High-Low Method (e.g., change in cost/activity).
   • For regression, state the equation and interpret coefficients.

3. Real-World Links:

   • Relate answers to Nepali businesses (e.g., "Like Daraz’s order processing costs").
   • Use local examples (e.g., Ncell’s subscriber-based costs).

4. Common Pitfalls:

   • High-Low Errors: Forgetting to use the highest/lowest points or miscalculating changes.
   • Regression Misuse: Assuming regression is always better (mention its limitations).
   • Units: Ensure costs are in Rs and activity in consistent units (e.g., units, hours).

Sample Exam Question: "A company’s electricity bill for 5 months is given below. Use the High-Low Method to segregate fixed and variable costs. Predict the bill for 2,500 units if the variable cost per unit is Rs 4."

Expected Answer Structure:

1. Identify highest/lowest activity levels.
2. Calculate variable cost per unit.
3. Derive fixed cost.
4. Formulate cost equation.
5. Plug in 2,500 units for prediction.

---

Visual Summary: