High Low Method vs. Regression Analysis

What is the high-low method and regression analysis? 

The high-low method and regression analysis are two common methods used in cost analysis to estimate the relationship between costs and activity levels. 

Both methods have benefits and drawbacks, and knowing how they differ can help organizations make more accurate and well-informed decisions.

The high-low method is a simple cost analysis technique that uses the highest and lowest data points to estimate fixed and variable costs. The method assumes that cost behavior is linear, meaning that costs vary proportionally with changes in activity levels. 

By calculating the cost difference between the greatest and lowest levels of activity and dividing that difference by the difference in activity levels, the high-low technique calculates the variable cost per unit of activity. 

By deducting the variable cost per unit from the total cost at either the high or low level of activity, the fixed cost can be determined once the variable cost has been identified. 

The high-low method is a useful technique for businesses that need to estimate costs quickly, as it only requires two data points and can be done manually or using spreadsheet software.

Regression analysis helps determine the relationship between two variables. The variables are usually the dependent variable, say X, and the independent variable, say Y. 

Regression analysis can be used in cost analysis to calculate the link between expenses and activity levels. It can also account for non-linear relationships between costs and activity levels. 

There are different types of regression analysis. The most prevalent types, among others, include non-linear regression, simple linear regression, and multiple regression.

The key differences between the two methods are presented below.

Usage of High-Low Method & Regression Analysis

When it comes to cost analysis, businesses often need to determine the relationship between costs and activity levels. 

The high-low method and regression analysis are two methods commonly used to estimate this relationship. An organization may face varied situations; some may use the high-low method, while sometimes, regression analysis will be preferred. 

1. High-Low Method

The high-low method is useful when a business needs to estimate fixed and variable costs quickly and does not have a large amount of data available. The method only requires two data points, the highest and lowest levels of activity, and assumes that cost behavior is linear. 

For example, there exists a small business by the hypothetical name ABC Traders. It receives a new order for some goods. In this situation, the business can deploy this method to determine a specific quantity of goods. 

This method, however, might not be appropriate for long-term cost analysis or when a company has a lot of data accessible. The method assumes that cost behavior is linear, which may not be accurate in all cases. 

2. Regression Analysis

When a company or organization is presented with a lot of data to be examined, regression analysis comes in handy. It takes into consideration non-linear relationships between costs and activity levels. 

Regression analysis can provide a more accurate estimate of fixed and variable costs by analyzing multiple data points and considering non-linear relationships. 

For example, XYZ Enterprises, a large manufacturing company, wants to estimate the cost of producing one of its bestseller products over the years. Consequently, it is presented with large chunks of data to be examined. 

In such a situation, XYZ Enterprises may use regression analysis to estimate the cost of producing a certain product over several years.

Regression analysis allows them to explore the relationship between the cost of production and various independent variables, such as time, raw material prices, labor hours, or any other relevant factors identified from the data.

However, this method may not be suitable when a business needs to make quick decisions based on limited data or when the cost behavior is linear. 

Additionally, this method requires more advanced statistical knowledge and may take more time.

Factors Affecting High-Low Method

The following factors affect the accuracy of the High-Low method.

1. Limited data points

The maximum and lowest levels of activity over a certain period are the only two data points needed for this strategy. The simplicity of this method is beneficial, but it opens room for errors because only limited data is used to produce results. 

2. Outliers

This method can produce results that are skewed by outliers or data points that are much higher or lower than the rest of the data. 

For example, if a company, say MN Corporation, experiences an unusual event, such as a strike or a natural disaster. This could significantly affect the results of this method.

3. Non-linear relationships

The high-low method assumes that the relationship between activity and cost is linear. This may not always happen in reality. For example, MN Corporation’s fixed costs may vary with changes in activity, such as changes in rent or insurance costs.

4. Seasonality

If a company experiences seasonal fluctuations in activity, this method may not accurately estimate fixed and variable costs. For example, if MN Corporation experiences a spike in activity during the holiday season, the high-low method may overestimate the fixed costs.

5. Changes in technology or production processes

Changes in technology or production processes can significantly affect cost behavior and render this method inaccurate. 

For example, if MN Corporation introduces a new technology that reduces labor costs, the high-low method may cause the method to overestimate the variable costs.

Factors Affecting Regression Analysis

The following factors affect the accuracy of the Regression Analysis.

1. Sample size

The precision of the findings in this method might be impacted by the sample size employed. Results are typically more accurate because a bigger sample size provides a more representative sample of the population being studied.

The outcome could not be representative of the entire population if the sample size is too small. Therefore, choosing an appropriate sample size that balances accuracy and precision is important.

2. Data quality

This method’s accuracy is based on how well the data are collected. The analysis's findings could be flawed if the data include inaccuracies or outliers. It is important to clean the data and remove any errors or outliers before conducting the regression analysis. 

Additionally, missing data can be handled through various other techniques, one of which is called imputation.

3. Multicollinearity

If an organization, say XYZ Corporation, performs regression analysis and notices a correlation between two or more variables, it will be said to have multicollinearity in its variables. 

Numerous diagnostic tests can be used to identify multicollinearity, and it can be treated using methods like principal component analysis or removing one of the associated variables.

4. Non-linearity

As we mentioned earlier, this method goes by the primary assumption that the relation between the taken dependent and independent variables is linear. However, this may not always happen, as the case may be. Non-linear relationships can lead to inaccurate results.

5. Overfitting

Overfitting occurs when a regression model is overly complex and fits the training data too well. This can lead to poor generalization and inaccurate predictions of new data. Overfitting can be addressed through techniques such as regularization or reducing the complexity of the model. 

High-Low Method Advantages

The high-low method has the following advantages.

1. Easy to use: People with a basic understanding of math can utilize this method since it is clear and simple. It only requires two data points, making it useful when data is limited or when a quick estimate is needed.
2. Useful for decision-making: This method can provide quick estimates of costs at different activity levels. This makes it useful for decision-making, especially when a deadline is involved.
3. Helps to understand cost behavior: This method can help businesses understand how costs change with different activity levels. By separating fixed and variable costs, businesses can better understand their cost structure and how to manage costs.
4. Can be used to estimate future costs: This method can be used to project future expenses at various activity levels in the future once the variable cost per unit of activity and the fixed cost have been estimated.

High-Low Method Disadvantages

High-Low Method has the following disadvantages:

1. Limited data points: The High-Low Method uses only two data points, which can result in less accurate cost estimates compared to methods that use more data points. Additionally, this method may not be appropriate for complex relationships between cost and activity.
2. Relies on linearity assumption: This method assumes that cost behavior is linear, meaning that costs vary proportionally with changes in activity levels. However, this assumption may not hold for all cost relationships, resulting in inaccurate cost estimates.
3. Ignores the effects of outliers: This method can be sensitive to outliers, which are extreme data points that may skew the results. Outliers can significantly impact the results, but the High-Low Method does not account for their effects.
4. Does not provide a detailed cost breakdown : While the High-Low Method can help businesses understand their cost structure, it does not provide a detailed breakdown of all costs. This can limit its usefulness in identifying cost drivers and managing costs.

Advantages & Disadvantages of Regression Analysis

The regression analysis has the following advantages.

1. Provides a quantitative analysis: This method provides a quantitative analysis of the relationship between variables. This helps in examining the data more precisely and objectively.
2. Identifies important factors: This method helps identify the important factors that are driving the relationship between the variables. This helps determine the elements that are most crucial for result prediction.
3. Forecasting: This method allows for the forecasting of future outcomes based on current data. This is useful in predicting trends and making informed decisions about future investments or business strategies.
4. Applicable to a wide range of fields: Regression analysis is a versatile method that can be applied in various disciplines including economics, finance, marketing, social sciences, and more.

On the other hand, it has the following disadvantages.

1. Requires assumptions: Certain relationships between the variables must be assumed to do a regression analysis, such as linearity, homoscedasticity, and normality. These presumptions could not always be accurate, which could result in unreliable findings.
2. Outliers can have a significant impact: This method is sensitive to outliers. This sensitivity significantly impacts the results of the analysis. Outliers can distort the relationship between the variables, leading to inaccurate predictions.
3. Requires a large sample size: This method requires a large sample size. This ensures that the results are statistically correct and of some significance. Small sample sizes can lead to inaccurate results.
4. Cannot establish causality: This method can establish a relationship between variables, but it cannot establish causality. Other factors may influence the relationship, and regression analysis cannot account for these factors.

High-Low Method vs. Regression Analysis: Practical Examples

Let’s take a look at an example of the High-low method.

A company named XYZ Corporation engaged in furniture production wants to estimate its monthly electricity bill based on its production level. The company's electricity bill for the last six months is as follows:

XYZ Corporation can deploy this method to estimate its variable cost per unit of furniture production.

Variable cost per unit = (Highest cost - Lowest cost) / (Highest production level - Lowest production level)

Variable cost per unit = $(2,700 - 1,200) / (600 - 100) = $6

XYZ Corporation can estimate its fixed cost by subtracting the variable cost from the total cost at either the high or low point. By using the high point (Month 6):

Fixed cost = Total cost - Variable cost * Production level

Fixed cost = $2,700 - $6 * 600 = $300

Therefore, the estimated monthly electricity bill can be calculated as

Monthly electricity bill = Fixed cost + Variable cost * Production level

Monthly electricity bill = $300 + $6 * Production level

Now, let’s take a look at an example of Regression Analysis.

The marketing team of XYZ Corporation wants to determine the relationship between the amount of money spent on advertising and the number of furniture units sold. They collect data for the last 12 months as follows:

Using this method, the marketing team can determine the linear relationship between advertising spending and units sold. They can use a statistical software package to run the regression analysis and obtain the following results

• Intercept (a): 39.212
• Slope (b): 0.011
• R-squared: 0.981

The regression equation can be written as

Units sold = 39.212 + 0.011 * Advertising spending

XYZ’s marketing team can make use of this obtained equation and estimate the number of units sold for any given amount of advertising spending. For example, if the marketing team plans to spend $17K on advertising next month, they can estimate the number of units sold as follows:

Units sold = 39.212 + 0.011 * $17,000 = 225.422

Researched and authored by Oorja Mahajan | Linkedin

Reviewed and Edited by Basil Khalidi | LinkedIn

Free Resources

To continue learning and advancing your career, check out these additional helpful WSO resources:

• Bullish and Bearish
• Business Life Cycle
• Buy-Side
• CAMELS rating system
• Calculating Foreign Exchange Spread