UNIT

09 Multiple Linear Regression

Names of Sub-Units

Introduction to multiple linear regression (MLR) analytical framework: Practical implementation of

MLR in Finance

Overview

In this unit you will study the concept of multiple linear regression (MLR) analytical framework.
Further, it discusses the practical implementation in the area of finance, utility of regression analysis,
assumptions of multiple linear regression analysis and MLR in financial modelling.

Learning Objectives

In this unit, you will learn to:

 Explain the introduction to multiple linear regression
 Assess the practical implementation in the area of finance
 Discuss the utility of regression analysis
 Analyse the assumptions of multiple linear regression analysis
 Appraise the MLR in financial modelling
JGI JAIN
DEEMED-TO-BE UNI VE RSI TY
Financial Analytics

Learning Outcomes

At the end of this unit, you would:

 Assess the introduction to multiple linear regression
 Appraise the practical implementation in the area of finance
 Evaluate the utility of regression analysis
 Discuss the assumptions of multiple linear regression analysis
 Explain the MLR in financial modelling

Pre-Unit Preparatory Material

 https://corporatefinanceinstitute.com/resources/knowledge/other/multiple-linear-regression/
 https://towardsdatascience.com/multiple-linear-regression-8cf3bee21d8b

9.1 INTRODUCTION
The term “regression” refers to statistical methods for fitting a model to data to quantify the relationship
between the dependent (outcome) variable and the predictor (independent) variable (s).

The dictionary meaning of the term ‘regression’ is the act of returning or going back. The term
‘regression’ was first used by Sir Francis Galton while studying the relationship between the heights
of fathers and sons. Regression analysis is the technique used for the modelling and analysis of
numerical data consisting of values of a dependent variable (response variable) and one or more
independent variables.

Now let us first try to understand what is the dependent and independent variable.
i. Dependent variable: It is the single variable being explained / predicted by the regression model
(Response variable).
ii. Independent variable: It is the explanatory variable used to predict the dependent variable
(Predictor variable).

The term ‘Regression’ is defined in multiple ways in the following definitions:

“Regression” is the measure of the average relationship between two or more variables in terms of the
original units of data.

According to Blair, “Regression is the measure of the average relationship between two or more variable
in terms of the original units of the data”.

According to Taro Yamane, ‘One of the most frequently used techniques in economics and business
research to find a relation between two or more variables that are related casually is regression analysis.’

2
UNIT 09: Multiple Linear Regression JGI JAIN
DEEMED-TO-BE UNI VE RSI TY

Utility of regression analysis:

a. Regression analysis helps in establishing a functional relationship between two or more variables.
Once this is established it can be used for various advanced analytical purposes.
b. Since most of the problems of economic analysis are based on cause-and-effect relationships, the
regression analysis is a highly valuable tool in economics and business research.
c. This can be used for the prediction or estimation of future production, prices, sales, investments,
income, profits and population which are indispensable for efficient planning of an economy and
are of paramount importance to a businessman or an economist.
d. Regression analysis is widely used in the statistical estimation of demand curves, supply curves,
production functions, cost functions, consumption functions, etc. Economists have discovered many
types of production functions by fitting regression lines to input and output data.

What are the ‘lines of regression’?

The line of regression is the line which gives the best estimate of the values of one variable for any
specific values of other variables.

For two variables on regression analysis, there are two regression lines. One line is the regression of X
on Y and the other for the regression of Y on X. These two regression lines show the average relationship
between the two variables. The regression line of Y on X gives the most probable values of Y for given
values of X and the regression line of X and Y gives the most probable values of X for given values of Y.

9.2 INTRODUCTION TO MULTIPLE LINEAR REGRESSION (MLR) ANALYTICAL FRAMEWORK

Multiple linear regression (MLR), often known as multiple regression, is a statistical technique that
predicts the result of a response variable by combining numerous explanatory variables. Multiple
linear regression attempts to represent the linear relationship between explanatory (independent) and
response (dependent) variables. Because it incorporates more than one explanatory variable, multiple
regression is essentially an extension of ordinary least-squares (OLS) regression.

Simple linear regression is a function that allows a statistician or analyst to generate predictions about
one variable based on data about another variable. Only two continuous variables—an independent
variable and a dependent variable—can be utilized in linear regression. The dependent variable or
outcome is calculated using the independent variable as a parameter. Multiple explanatory variables
are included in a multiple regression model.

The coefficient of determination (R-squared) is a statistical metric that measures how much change in
the independent variables can be explained by variance in the result. Even if the predictors are unrelated
to the outcome variable, R2 grows as more predictors are added to the MLR model.

As a result, R2 alone cannot be used to determine which predictors should be included and which
should be eliminated from a model. R2 can only be between 0 and 1, with 0 indicating that none of the
independent variables can predict the outcome and 1 indicating that the independent variables can
predict the outcome without mistake.

3
JGI JAIN
DEEMED-TO-BE UNI VE RSI TY
Financial Analytics

While holding all other variables constant, beta coefficients are appropriate when evaluating the results
of multiple regression (“all else equal”). A multiple regression’s outcomes might be shown horizontally
as an equation or vertically as a table.

There are various types of regression and those are – simple regression, multiple regression, curvilinear
regression and polynomial regression.

The multiple regression analysis represents a logical extension of two-variable regression analysis.
Instead of a single independent variable, two or more independent variables are used to estimate the
values of a dependent variable. However, the fundamental concepts in the analysis remain the same.

Multiple linear regression (MLR), often known as multiple regression, is a statistical technique that
predicts the result of a response variable by combining numerous explanatory variables. Multiple
linear regression attempts to represent the linear relationship between explanatory (independent) and
response (dependent) variables. Because it incorporates more than one explanatory variable, multiple
regression is an extension of ordinary least-squares (OLS) regression.

The link between two or more independent variables and one dependent variable is estimated using
multiple linear regressions. This multiple linear regression can be used when one wants to know:
1. The degree to which two or more independent variables and one dependent variable are related (i.e.,
how rainfall, temperature and amount of fertilizer added affect crop growth)
2. The dependent variable’s value at a given value of the independent variables (i.e., the expected yield
of a crop at certain levels of rainfall, temperature and fertilizer addition)

In multiple regression analysis, there are three or more variables say X1, X2 and X3. We now take X1 as the
dependent variable and try to find out its relative movement for movements in both X2 and X3, which
are independent variables. Thus, in multiple regression analysis, the effect of two or more independent
variables on one dependent variable is studied.

If there are three variables X1, X2 and X3, the multiple regressions will take the following form:
X1 = a1.23 + b12.3 X2 + b13.2 X3

It can be noted that in the above equation there are three constraints. The subscript after the dot
indicates the variables, which are held constant.

If X2 and X3 are 0 then X1 = a1.23

For multiple regression models with two independent variables, the result is three simultaneous
equations with three unknowns (b0, b1 and b2).

We can now note the formula for calculation of Multiple Linear Regression as follows:
yi = B0 + B1xi1 + B2xi2 + …… + Bpxip + £

Where,
For i = n observations
yi = dependent variable

4
UNIT 09: Multiple Linear Regression JGI JAIN
DEEMED-TO-BE UNI VE RSI TY

xi = explanatory variables
B0 = y – intercept (constant term)
Bp = slope coefficients for each explanatory variable
£ = the model’s error term (also known as residuals)

Multiple regression models generally involve lengthy calculations and hence they are handled by
statistical software’s like SPSS, MINITAB, MS-EXCEL etc.

Assumptions of Multiple Linear Regression Analysis

For point estimation, the principle assumptions of linear multiple regression analysis are:
1. The dependent variable is a random variable whereas the independent variables need not be
random variable.
2. The relationship between the several independent variables and the one dependent variable is
linear.
3. The variances of the conditional distributions of the dependent variable given various combinations
of values of the independent variables are all equal.

For internal estimation, an additional assumption is that the conditional distributions for the dependent
variable follows the normal probability distribution.

9.3 PRACTICAL IMPLEMENTATION OF MLR IN FINANCE

As we have seen in the beginning stage of this unit, in business, regression analysis is a statistical
technique for determining the relationships between two or more independent and dependent variables.
The MLR analytical framework proves very useful in the business environment as well financial decision
making and policy framing.

One of the most often used statistical techniques is linear regression. It is used to measure the link
between a response variable and one or more predictor factors.

Simple linear regression is the most basic type of linear regression and it’s used to quantify the
relationship between one predictor variable and one response variable.

We can use multiple linear regression to quantify the link between numerous predictor variables and a
response variable if we have more than one predictor variable.

This course shows four diverse applications of linear regression in the real world.

Linear Regression Real Life Example #1

To understand the relationship between advertising spending and income, businesses frequently employ
linear regression.

5
JGI JAIN
DEEMED-TO-BE UNI VE RSI TY
Financial Analytics

They may run a basic linear regression model with advertising spending as the predictor variable and
revenue as the response variable, for example. The regression model would look something like this:

Revenue = 0 + 1 (ad spending)

When ad spending is nil, the coefficient 0 represents the entire expected revenue.

When ad spending is increased by one unit, coefficient 1 represents the average change in total revenue
(i.e., one dollar).

If 1 is negative, it means that more advertising spending equals less revenue.

If 1 is close to 0, advertising expenditure has little impact on income.

If 1 is positive, it means that greater advertising spending equals more income.

A company’s advertising spending may be reduced or increased depending on the value of 1.

Linear Regression Real Life Example #2

Linear regression is frequently used by medical researchers to better understand the association
between drug dosage and patient blood pressure.

For example, researchers might give patients different doses of medicine and watch how their blood
pressure changes. They might use dosage as the predictor variable and blood pressure as the response
variable in a simple linear regression model. The regression model would look something like this:

Blood pressure = 0 + 1 (dosage)

When the dosage is zero, the coefficient 0 represents the predicted blood pressure.

When the dosage is increased by one unit, coefficient 1 represents the average change in blood pressure.

If 1 is negative, it means that increasing the dosage leads to a reduction in blood pressure.

If 1 is near zero, an increase in dosage is not connected with an increase in blood pressure.

If 1 is positive, it means that increasing the dosage leads to an increase in blood pressure.

Researchers may opt to adjust the dosage provided to a patient based on the value of 1.

Linear Regression Real Life Example #3

Agricultural scientists frequently employ linear regression to determine how fertiliser and water affect
crop yields.

For instance, scientists may experiment with varing amounts of fertiliser and water in different areas
to examine how crop productivity is affected. They might use fertiliser and water as predictor factors
and crop yield as the response variable in a multivariate linear regression model. The regression model
would look something like this:

Crop yield = 0 + 1 (amount of fertilizer) + 2 (amount of water)

6
UNIT 09: Multiple Linear Regression JGI JAIN
DEEMED-TO-BE UNI VE RSI TY

With no fertiliser or water, coefficient 0 represents the projected crop output.

If the amount of water is kept constant, the coefficient 1 represents the average change in crop output
when fertiliser is increased by one unit.

If the amount of fertiliser is kept constant, the coefficient 2 represents the average change in crop
output when water is increased by one unit.

Scientists may adjust the amount of fertiliser and water used to enhance crop yield based on the values
of 1 and 2.

Linear Regression Real Life Example #4

Professional sports teams frequently utilize linear regression to assess the impact of different training
regimens on player performance.

For example, in the NBA, data scientists might look at how varied quantities of weekly yoga and
weightlifting sessions affect a player’s point total. They might use yoga and weightlifting sessions as
predictor factors and total points achieved as the response variable in a multivariate linear regression
model. The regression model would look something like this:

Points scored = 0 + 1(yoga sessions) + 2(weightlifting sessions)

The coefficient 0 represents the predicted points for a player who does not participate in any yoga or
weightlifting workouts.

If the number of weekly weightlifting sessions remains constant, the coefficient 1 represents the average
change in points obtained when weekly yoga sessions are increased by one.

When weekly weightlifting sessions are increased by one, the coefficient 2 represents the average
change in points scored, assuming the number of weekly yoga sessions remains constant.

To optimise points collected, data scientists may recommend that a player participates in more or less
weekly yoga and weightlifting sessions, depending on the values of 1 and 2.

Following are the various useful applications of multiple linear regressions in finance as well as the
business as a whole. Let us now try to understand those applications in detail:
1. Beta and CAPM: There are several applications of regression in finance. The Capital Asset Pricing
Model (CAPM) relies heavily on statistical methods. The CAPM equation is essentially a formula that
determines the relationship between an asset’s expected return and the market risk premium. The
technique can also be used to anticipate the returns of securities depending on a variety of criteria,
or to forecast a company’s success.
Regression analysis is used in finance to calculate a stock’s Beta (volatility of returns relative to the
general market). The Slope function in Excel can be used to accomplish this.
2. Forecasting revenue and expenses: When forecasting financial accounts for a firm, a multiple
regression analysis may be useful to see how changes in particular assumptions or business
drivers may affect revenue or expenses in the future. For example, there may be a strong link

7
JGI JAIN
DEEMED-TO-BE UNI VE RSI TY
Financial Analytics

between a company’s number of salespeople, the number of stores it operates and the amount of
income it makes.
3. Predictive analytics: The most common application of regression analysis in business is predictive
analytics, which involves projecting future opportunities and hazards. Demand analysis, for
example, forecasts the number of things a customer is likely to buy. When it comes to business,
however, demand is not the only dependent variable.
4. Operation efficiency: Business processes can also benefit from regression models. For example,
a plant manager can develop a statistical model to determine the effect of oven temperature on
the shelf life of cookies made in those ovens. Guesswork, supposition and corporate politics are all
eliminated in data-driven decision making. This enhances corporate performance by focusing on
the areas that have the greatest impact on operational efficiency and revenue.
5. Decision support: Today’s businesses are inundated with information about their finances,
operations and customer transactions. Executives are increasingly relying on data analytics to
make informed business decisions with statistical significance, removing the need for intuition
and gut feeling.
6. Correction of errors: Regression is useful not just for providing empirical evidence for management
decisions, but also for detecting judgement errors. A retail store manager, for example, may assume
that extending shopping hours will significantly boost sales. However, RA could suggest that the
increase in revenue isn’t enough to cover the increase in operating expenses caused by longest
working hours.

MLR in Financial Modelling

In most real-world scenarios, simple regression is insufficient since objectives (dependent variables) are
rarely influenced by a single predictor. The sole difference between the multiple linear regression model
and the simple one is that it can have two or more independent variables (predictors). The following
equation represents the regression function:
Y = a + b.x1 + c.x2 + d.x3 + £

It’s important to remember that the multiple regression model necessitates non-collinearity. This implies
that the independent variables should have a low correlation. The linear regression model is used in the
Capital Asset Pricing Model (CAPM), which determines the relationship between an asset’s predicted
return and the associated market risk premium. Financial analysts frequently use it to anticipate
corporate returns and operational performance.

The beta coefficient of a stock is calculated using regression analysis. The volatility of returns is
displayed in relation to the overall market risk. The SLOPE function in Excel can quickly calculate it since
it represents the slope of the CAPM regression.

We may choose to run a multiple linear regression analysis when forecasting financial statements as
part of our budgeting and planning exercises to see how our model assumptions would affect future
business performance. Sales forecasting is one of the most common applications of regression analysis.

8
UNIT 09: Multiple Linear Regression JGI JAIN
DEEMED-TO-BE UNI VE RSI TY

Conclusion 9.4 CONCLUSION

 The term “regression” refers to statistical methods for fitting a model to data in order to quantify the
relationship between the dependent (outcome) variable and the predictor (independent) variable (s).
 Dependent variable: it is the single variable being explained / predicted by the regression model
(response variable).
 Independent variable: it is the explanatory variable used to predict the dependent variable (predictor
variable).
 “Regression” is the measure of the average relationship between two or more variables in terms of
the original units of data.
 The line of regression is the line which gives the best estimate of the values of one variable for any
specific values of other variables.
 There are various types of regression and those are – simple regression, multiple regression,
curvilinear regression and polynomial regression.
 The multiple regression analysis represents a logical extension of two-variable regression analysis.
 Multiple linear regression (MLR), often known as multiple regression, is a statistical technique that
predicts the result of a response variable by combining numerous explanatory variables.
 The link between two or more independent variables and one dependent variable is estimated using
multiple linear regression.
 Following are the various useful applications of multiple linear regression in finance as well as the
business as a whole:
i. Beta and CAPM
ii. Forecasting revenue and expenses
iii. Predictive analytics
iv. Operation efficiency
v. Decision support
vi. Correction of errors
 In most real-world scenarios, simple regression is insufficient since objectives (dependent variables)
are rarely influenced by a single predictor. The sole difference between the multiple linear regression
model and the simple one is that it can have two or more independent variables (predictors).
 The linear regression model is used in the Capital Asset Pricing Model (CAPM), which determines the
relationship between an asset’s predicted return and the associated market risk premium.
 The beta coefficient of a stock is calculated using regression analysis. The volatility of returns is
displayed in relation to the overall market risk.
 Sales forecasting is one of the most common applications of regression analysis.

9
JGI JAIN
DEEMED-TO-BE UNI VE RSI TY
Financial Analytics

9.5 GLOSSARY

 Regression: The term “regression” refers to statistical methods for fitting a model to data to quantify
the relationship between the dependent (outcome) variable and the predictor (independent)
variable (s).
 Capital Asset Pricing Model (CAPM): The CAPM equation is essentially a formula that determines
the relationship between an asset’s expected return and the market risk premium.
 Forecasting: When forecasting financial accounts for a firm, a multiple regression analysis may
be useful to see how changes in particular assumptions or business drivers may affect revenue or
expenses in the future.
 Operation efficiency: Business processes can also benefit from regression models. For example, a
plant manager can develop a statistical model to determine the effect of oven temperature on the
shelf life of cookies made in those ovens.

9.6 CASE STUDY: REGRESSION TREE

Case Objective
This case study highlights the regression tree outperformed multiple linear regression in predicting
the relationship between impairments and Social and Productive Activities scores
 Objective: There are numerous key physiologic and clinical predictors that remain constant. The
interest in these predictors among clinical investigators and epidemiologists stems in part from the
likelihood of bad outcomes, which may be constant. The relationship between continuous predictors
and continuous outcomes might be tricky to decipher. Clinical investigators and epidemiologists
might benefit from methods to detect levels of a predictor variable that predict the result and define
the threshold for clinical intervention.
 Study design and setting: We offer a case study in which we used regression tree methods to predict
the Social and Productive Activities score at three years utilising five modifiable impairments.
Using two separate data sets, one for development and one for validation, the predictive ability of
the regression tree methodology was compared to multiple linear regression.
 Results: We provide a case study in which five modifiable impairments were utilized to predict
the Social and Productive Activities score at three years using regression tree approaches. The
predictive ability of the regression tree approach was compared to multiple linear regression using
two independent data sets, one for development and one for validation.
 Conclusion: We provide a case study in which five modifiable impairments were utilized to predict
the Social and Productive Activities score at three years using regression tree approaches. The
predictive ability of the regression tree approach was compared to multiple linear regression using
two independent data sets, one for development and one for validation.

10
UNIT 09: Multiple Linear Regression JGI JAIN
DEEMED-TO-BE UNI VE RSI TY

Questions
1. Discuss the objective of this case study.
(Hint: There are numerous key physiologic and clinical predictors that remain constant. The
interest in these predictors among clinical investigators and epidemiologists stems in part from the
likelihood of bad outcomes, which may be constant.)
2. Discuss the conclusion of this case study.
(Hint: The predictive ability of the regression tree approach was compared to multiple linear
regression using two independent data sets, one for development and one for validation.)

9.7 SELF-ASSESSMENT QUESTIONS

A. Essay Type Questions:

1. What do you mean by the term ‘Regression’?
2. What do you mean by Multiple Linear Regression? Also, state the equation of MLR.
3. What are the various useful applications of Multiple Linear Regressions in finance and business as
a whole?
4. Explain how to apply MLR in finance?

9.8 ANSWERS AND HINTS FOR SELF-ASSESSMENT QUESTIONS

A. Hints for Essay Type Questions

1. The term “regression” refers to statistical methods for fitting a model to data in order to quantify the
relationship between the dependent (outcome) variable and the predictor (independent) variable (s).
The term ‘Regression’ is defined in multiple ways in the following definitions:
Refer to Section of Introduction to Multiple Linear Regression
2. Multiple linear regression (MLR), often known as multiple regression, is a statistical technique that
predicts the result of a response variable by combining numerous explanatory variables. Multiple
linear regression attempts to represent the linear relationship between explanatory (independent)
and response (dependent) variables. Because it incorporates more than one explanatory variable,
multiple regression is an extension of ordinary least-squares (OLS) regression.
If there are three variables X1, X2 and X3, the multiple regression will take the following form:
X1 = a1.23 + b12.3 X2 + b13.2 X3
It can be noted that in above equation there are three constraints. The subscript after the dot
indicates the variables, which are held constant.
If X2 and X3 are 0 then X1 = a1.23
Refer to Section of Introduction to Multiple Linear Regression

11
JGI JAINDEEMED-TO-BE UNI VE RSI TY
Financial Analytics

3. The MLR analytical framework proves very useful in the business environment as well financial
decision making and policy framing. Following are the various useful applications of multiple linear
regression in finance as well as the business as a whole:
i. Beta and CAPM
ii. Forecasting revenue and expenses
iii. Predictive analytics
iv. Operation efficiency
v. Decision support
vi. Correction of errors
Refer to Section of Practical Implementation in the Area of Finance
4. It’s important to remember that the multiple regression model necessitates non-collinearity. This
implies that the independent variables should have a low correlation. The Capital Asset Pricing
Model (CAPM), which determines the relationship between an asset’s projected return and the
related market risk premium, relies on the linear regression model. Financial analysts frequently
use it to anticipate corporate returns and operational performance.
Refer to Section of Practical Implementation in the Area of Finance

@ 9.9 POST-UNIT READING MATERIAL

 https://statsandr.com/blog/multiple-linear-regression-made-simple/
 https://www.voxco.com/blog/multiple-regression-analysis-definition-example-and-equation/

9.10 TOPICS FOR DISCUSSION FORUMS

 Make a group of three/four friends to discuss the concept of multiple linear regression.

12