Demand Theory and Estimation: Detailed Notes

Demand theory explores the relationship between the price of a product

or service and the quantity demanded by consumers. For firms,
understanding demand is crucial for making pricing, production, and
marketing decisions. Demand estimation, on the other hand, involves
using statistical techniques to forecast demand based on historical data.

1. Demand Faced by a Firm

For a firm, the demand curve is the relationship between the price it
charges for a product and the quantity of that product consumers are
willing to buy at various price levels. This relationship is influenced by
factors such as consumer preferences, income levels, and the prices of
substitutes and complements.
Key Concepts:
 Price Elasticity of Demand: Measures the responsiveness of
quantity demanded to a change in price. It is calculated as:
% change∈quantity demanded price
Ed= % change∈ price . quantity demanded

 Income Elasticity of Demand: Measures the responsiveness of

quantity demanded to changes in consumer income.
 Cross-Price Elasticity of Demand: Measures how the quantity
demanded of one good responds to changes in the price of another
good.

2. Techniques of Demand Estimation

Demand estimation involves statistical and econometric techniques to
model and predict demand. Common methods include:
  Survey Methods: Gathering data directly from potential
consumers through surveys.
 Market Experiments: Testing consumer reactions to different
prices in a controlled environment.
 Regression Analysis: A statistical technique used to estimate the
relationship between demand and its determinants.
Regression Analysis for Demand Estimation
Regression analysis is one of the most commonly used methods for
demand estimation. It allows firms to quantify the relationship between
quantity demanded and various independent variables (e.g., price,
income, advertising).
Example of a Demand Function
A simple linear demand function can be represented as:
Qd=a+bP+cI+dA
where:
 Qd = Quantity demanded
 P = Price of the product
 I = Consumer income
 A = Advertising expenditure
 a,b,c, and d = Coefficients to be estimated using regression
analysis
Price, income, Qd=a-bP+cI+dA
advertising

Quantity demanded

Regression Method Example with Mathematical Calculations

Let’s estimate the demand for a product using Ordinary Least Squares
(OLS) regression. Suppose a firm collects data on the quantity
demanded, price, income, and advertising expenditure over a period.
Step-by-Step Process
1. Collect Data: Assume the following data on Quantity (Q), Price
(P), Income (I), and Advertising (A).
Period Quantity (Q) Price (P) Income (I) Advertising (A)
1 100 20 5000 300
2 120 18 5200 320
3 90 22 5100 310
4 110 19 5300 330
5 95 21 5000 315
2. Formulate the Demand Equation: Assume a linear demand
model as follows:
Q=a+bP+cI+dA+e
where e is the error term.
3. Run the Regression Analysis:
Using software like Excel, Stata, or R, perform regression analysis to
estimate the coefficients a,b,c,and d.
4. Interpret the Results:
5.
Coefficientsa
Standardized
Unstandardized Coefficients Coefficients
Model B Std. Error Beta t Sig.
1 (Constant) 191.005 110.229 1.733 .333
PRICE -6.984 1.420 -.917 -4.919 .128
INCOME .008 .029 .092 .291 .820
ADVERTISING .026 .310 .025 .085 .946
a. Dependent Variable: QUANTITY

6.

Suppose the estimated regression equation is:

Q=191.005+(−6)P+0.008I+0.26A
This means:
o The intercept a=191.005: Baseline quantity demanded when
all independent variables are zero.
o b=−6: For every 1-unit increase in price, quantity demanded
decreases by 6 units.
o c=0.008: For every 1-unit increase in income, quantity
demanded increases by 0.008units.
o d=0.26: For every 1-unit increase in advertising expenditure,
quantity demanded increases by 0.26 units.
Forecasting Demand
Using the estimated coefficients, the firm can forecast demand. For
instance, if:
 Price P=20
 Income I=5100
 Advertising A=320
Then,
Q=191.005+(−6⋅20)+(0.008⋅5100)+(0.26⋅320)
Breaking it down:
1. −6⋅20=−120
2. 0.008⋅5100=40.8
3. 0.26⋅320=83.2
So,
Q=191.005−120+40.8+83.2=195.005
Thus, the estimated quantity demanded is 195.005 units under the given
conditions.

3. Techniques of Demand Estimation (Continued)

A. Survey Method (Detailed)
 Designing the Survey: Create a questionnaire that includes questions about consumer
preferences, price sensitivity, and purchasing habits.
 Sampling: Select a representative sample of the target market to ensure the results are
generalizable.
 Data Collection: Administer the survey through various channels (online, in-person,
etc.) and collect responses.
 Analysis: Analyze the data to identify trends and preferences that can inform demand
estimates.
B. Market Experiments (Detailed)
 Experimental Design: Choose a specific market or demographic to test different price
points or marketing strategies.
 Implementation: Change the price or marketing approach in the selected market while
keeping other factors constant.
 Data Collection: Monitor sales and consumer responses during the experiment.
 Analysis: Compare the results with a control group to assess the impact of the changes
on demand.
C. Statistical Methods (Detailed)
 Regression Analysis: Use historical sales data to create a regression model that predicts
demand based on various independent variables (price, income, etc.).
 Model Validation: Test the model with a separate dataset to ensure its predictive
accuracy.
 Forecasting: Use the validated model to forecast future demand under different
scenarios.

REGRESSION ANALYSIS
To estimate the regression coefficients in a simple linear regression
model, we can use formulas derived from the Ordinary Least Squares
(OLS) method. The OLS method minimizes the sum of the squared
differences between observed values and predicted values.
Let’s go through the process of calculating the coefficients aaa and bbb
in a linear regression model:
Simple Linear Regression Model
In a simple linear regression, we have:
Y=a+bX+e
where:
 Y is the dependent variable,
 X is the independent variable,
  a is the intercept (or constant term),
 b is the slope coefficient, and
 e is the error term.
Formulas to Estimate a and b
The formulas to estimate a and b (known as the least squares
estimators) are:
∑( Xi− Xˉ )(Yi−Yˉ )
b¿ ∑ (Xi−Xˉ )2

Where:
 Xi and Yi are individual observed values,
 Xˉ is the mean of the X values, and
 Yˉ is the mean of the Y values.
Step-by-Step Calculation
1. Calculate Xˉ\bar{X}Xˉ and Yˉ\bar{Y}Yˉ:
∑ Xi ∑ Yi
Xˉ= n ,Yˉ= n

where n is the number of observations.

2. Calculate the Slope b: Substitute the values into the formula for b:
∑( Xi− Xˉ )(Yi−Yˉ )
b¿ ∑ (Xi−Xˉ )2

3. Calculate the Intercept a: Using the calculated value of b and the

means Xˉ and Yˉ, we can find aa:
a=Yˉ−bXˉ
Example Calculation
Let’s say we have the following data points:
X (Independent Variable) Y (Dependent Variable)
1 2
2 3
3 5
4 7
5 8
Step 1: Calculate Xˉ\bar{X}Xˉ and Yˉ\bar{Y}Yˉ
1+ 2+ 3+4 +5
Xˉ= 5 =3
2+ 3+5+7+ 8
Yˉ= 5 =5

Step 2: Calculate the Slope b

We can now calculate b using the formula:
∑( Xi− Xˉ )(Yi−Yˉ )
b¿ ∑ (Xi−Xˉ )2

First, find (Xi−Xˉ)( and (Yi−Yˉ) for each data point and their products.
Xi Yi Xi−Xˉ=X* Yi−Yˉ= Y* (Xi−Xˉ)(Yi−Yˉ) (Xi−Xˉ)2
1 2 -2 -3 6 4
2 3 -1 -2 2 1
Xi Yi Xi−Xˉ=X* Yi−Yˉ= Y* (Xi−Xˉ)(Yi−Yˉ) (Xi−Xˉ)2
3 5 0 0 0 0
4 7 1 2 2 1
5 8 2 3 6 4
∑ X∗¿=6 ¿ ∑ Y ∗¿=10 ¿ ∑ X∗Y ∗¿= 16 ∑ X∗2=10

n=5
Now, sum up the values in the last two columns:
∑ (Xi−Xˉ) (Yi−Yˉ) =6+2+0+2+6=16
∑ (Xi−Xˉ) 2=4+1+0+1+4=10
∑( Xi− Xˉ )(Yi−Yˉ )
Now, calculate b: ¿ ∑ (Xi−Xˉ )2

16
b= 10 =1.6

Step 3: Calculate the Intercept aaa

Using the formula a=Yˉ−bXˉ
a=5−(1.6⋅3)=5−4.8=0.2

Final Regression Equation

The estimated regression equation is:
Y=0.2+1.6X
This equation implies that for every 1-unit increase in X, Y increases by
1.6 units.
ANOTHER FORMULA TO BE USED IN CALCULATING
REGRESSION EQUATIONS
∑Y*= na+b∑ x∗… … … … … … … … … i
∑X*Y*= a∑ x∗+ b ∑ x∗2………………….ii

After the equation solve using either substitution method and elimination
method.
10= 5a+6b………………..i
16= 6a+10b……………..ii
16 -10 b= 6a
16 10
− =a
6 6
8 5
a= 3 - 3 b
5
10=5( ¿ ¿- 3 b ¿+ 6 b
1 1
10= 13 3 -8 3 b +6b

10= 40/3-7/3b
10-40/3=
-10/3= -7/3b
b= 10/7= 1.428571429
a= 8/3-5/3(10/7)
a= 0.285714285714286
Y= 0.285714285714286+1.428571429X
The expression in fraction form becomes:

2 10
y= + X
7 7
Let’s estimate y at x= 6
 2 10
y= + (6)
7 7

1. Calculate the multiplication first:

10 60
×6=
7 7

Now substitute this back into the equation:

2 60
y= +
7 7

2. Combine the fractions:

2+ 60 62
y= =
7 7

3. So, the value of y is:

y≈8.8571

. The differences in solutions obtained from the two methods—the formula method and the
simultaneous equations method—may arise from several factors, including numerical precision,
rounding errors, or the specific assumptions made in each method. Here’s a breakdown of each
approach:

Formula Method

1. Derived from Sums: The formula method directly utilizes sums of products and squares
from the data points. It calculates aaa and bbb based on aggregate statistics derived from
the data.
2. Least Squares Estimation: This method employs least squares estimation principles,
aiming to minimize the sum of squared errors between observed and predicted values.
Thus, it is closely tied to the statistical properties of the data set.
3. Impact of Data Variability: Variability in data can influence results. If data points are
outliers, the calculated regression coefficients may differ when based on different aspects
of the data set (e.g., total sums versus individual equations).

Simultaneous Equations Method

1. Set of Linear Equations: This method uses derived linear equations based on aggregate
sums to create a system of simultaneous equations.
2. Dependency on Variables: The results depend on how the equations are structured and
solved. Any slight changes in approach to forming these equations or in calculating
values can lead to variations in the results.
 3. Solving Technique: The method of solving (whether substitution or elimination) may
introduce discrepancies. For example, if rounding occurs during calculations, this can
lead to different results.

Reasons for Discrepancies

1. Rounding Errors: In both methods, numerical rounding at various steps can accumulate
and lead to differences in the final estimates for aaa and bbb.
2. Data Sensitivity: Regression coefficients can be sensitive to the specific values used,
particularly if the data set has outliers or is not perfectly linear.
3. Assumptions in Calculation: Each method might be making different assumptions about
the data or how the relationships between variables should be modeled.

Conclusion

In practice, if both methods are correctly applied and yield significantly different coefficients, it
is essential to check each calculation step for accuracy and consistency. It might also be
beneficial to use statistical software to verify results, as these tools will typically handle rounding
and numerical precision issues more effectively.

For further reading on regression analysis and the methods used, you might find useful insights
in resources such as:

 Statistics for Business and Economics

 Introduction to Linear Regression Analysis

D. Time Series Analysis (Detailed)

 Data Collection: Gather historical sales data over a significant period.
 Trend Analysis: Identify trends, seasonal patterns, and cyclical movements in the data.
 Forecasting Models: Use models such as ARIMA (AutoRegressive Integrated Moving
Average) to predict future demand based on past patterns.
E. Econometric Models (Detailed)
 Complex Models: Develop econometric models that incorporate multiple variables
affecting demand, such as advertising expenditure, consumer demographics, and
economic indicators.
 Estimation Techniques: Use techniques like Maximum Likelihood Estimation (MLE) or
Generalized Method of Moments (GMM) to estimate the parameters of the model.
  Policy Analysis: Use the model to simulate the effects of policy changes or market
conditions on demand.
Conclusion
Understanding demand theory and estimation is crucial for firms to make informed decisions
regarding pricing, production, and marketing strategies. By employing various techniques for
demand estimation, businesses can better predict consumer behavior and adapt to changing
market conditions. The mathematical calculations involved in demand estimation, such as
regression analysis and elasticity calculations, provide valuable insights that can guide strategic
planning and operational efficiency.
These notes provide a comprehensive overview of demand theory and estimation, including
practical applications and mathematical calculations that are essential for understanding how
demand functions in the marketplace
Risk Analysis
Risk analysis in business and economics is a critical process for
identifying and evaluating factors that could negatively impact key
initiatives or projects. It involves assessing potential risks, their
probabilities, and the impact they may have on decision-making and
outcomes. Here’s an in-depth look at the key concepts in risk analysis,
including risk and probability, the nature of risk and uncertainty in
managerial decisions, and decision rules and valuation models for
assessing risk.

1. Risk and Probability

Risk and probability are foundational concepts in risk analysis. Risk is
the exposure to a situation involving the possibility of loss or gain, while
probability is the measure of the likelihood that a particular event will
occur.
 Definition of Risk: In the context of business decisions, risk
represents potential variability in outcomes due to uncertain
factors. Risks are often classified into various types:
o Financial Risks: Relate to market fluctuations, currency
rates, and credit risks.
o Operational Risks: Arise from internal processes, systems,
and people.
o Strategic Risks: Linked to external factors like competition,
technological changes, and regulatory shifts.
 Definition of Probability: Probability quantifies the likelihood of
a certain event happening, expressed as a value between 0
(impossible) and 1 (certain). Probabilities can be calculated
objectively using historical data or estimated subjectively based on
judgment.
Risk Assessment Using Probability
The probability of an event helps in understanding and measuring risk:
 Expected Value (EV): This is a basic measure in risk analysis,
calculated as:
Expected Value=∑(Probability of Outcome×Value of Outcome)
 Variance and Standard Deviation: Measures of the spread of
possible outcomes around the expected value, giving an idea of the
risk involved.
 Probability Distributions: Different probability distributions, like
the normal distribution or Poisson distribution, model various
types of risks and their outcomes.

2. Risk and Uncertainty in Managerial Decisions

In the business world, managers frequently make decisions under
conditions of risk and uncertainty. These concepts are interrelated but
differ in how they are handled:
 Risk: Exists when decision-makers can assign probabilities to
outcomes based on past data or theoretical understanding. For
example, insurance companies calculate premiums based on
statistical risk data.
 Uncertainty: Arises when probabilities cannot be determined,
often due to a lack of data or unpredictable conditions. For
example, a company entering a completely new market faces
uncertainty since there is no prior information.
Types of Uncertainty in Managerial Decisions
1. Decision Uncertainty: Due to incomplete information about the
choices available.
 2. State Uncertainty: Occurs when the environment or external
conditions are unpredictable.
3. Effect Uncertainty: Involves uncertainty about the impact or
consequences of a decision.
Strategies for Managing Uncertainty
Managers can address uncertainty in decisions by:
 Scenario Planning: Developing multiple scenarios to understand
the range of possible outcomes.
 Flexibility: Incorporating adaptability in business plans and
projects.
 Information Gathering: Increasing the amount of relevant
information to reduce uncertainty.
 Probabilistic Forecasting: Estimating the likelihood of different
future events to make informed predictions.

3. Decision Rule and Valuation Model of Risk

Decision rules and valuation models provide systematic approaches to
make decisions under conditions of risk.
Decision Rules
Decision rules are criteria or formulas used to guide choices among risky
alternatives. Common decision rules include:
1. Maximin Rule: Choose the alternative with the best worst-case
outcome. This approach is conservative and suitable for risk-averse
individuals.
2. Maximax Rule: Choose the alternative with the best possible
outcome, regardless of risk. It’s often used by risk-takers who
prioritize potential maximum gains.
 3. Minimax Regret Rule: Focuses on minimizing regret by choosing
an option that limits potential loss relative to the best alternative.
4. Expected Value Maximization: Involves selecting the alternative
with the highest expected value, balancing risk and return. This
approach assumes a risk-neutral perspective.
Valuation Models of Risk
Valuation models help quantify risk and assess its impact on potential
decisions:
1. Expected Monetary Value (EMV):
o Uses probability-weighted averages to compute the financial
outcome of different alternatives. The formula is:
EMV=∑(Probability of Outcome×Monetary Value of Outco
me)
o Example: If there’s a 40% chance of earning $10,000 and a
60% chance of earning $4,000, the EMV would be:
EMV=(0.4×10,000)+(0.6×4,000)=4,000+2,400=6,400
2. Variance and Standard Deviation:
o Used to measure the spread or volatility of outcomes around
the mean or expected value. Higher variance indicates higher
risk.
o Example: For the outcomes above, if the variance is
calculated, a high variance indicates a wider range of
possible outcomes.
3. Decision Trees:
o A visual representation of different decision paths and their
possible outcomes, including probabilities and payoffs.
Decision trees are used in complex decisions where
sequential choices are involved.
 4. Sensitivity Analysis:
o Analyzes how sensitive outcomes are to changes in input
variables. It is useful for understanding which factors have
the most influence on risk.
5. Monte Carlo Simulation:
o Uses random sampling and statistical modeling to estimate
the impact of risk. This approach is valuable when dealing
with multiple variables and complex risk structures.
6. Discounted Cash Flow (DCF) Analysis:
o Evaluates the present value of expected future cash flows to
account for time and risk. The Net Present Value (NPV)
formula, which discounts future cash flows at a chosen
Cash Flow t
discount rate, is: NPV=∑ (1+r )t −Initial Investment

o The choice of discount rate reflects the riskiness of the cash

flows. Higher risk leads to a higher discount rate, reducing
the NPV.
7. Value at Risk (VaR):
o Measures the maximum potential loss over a given time
period, with a specific confidence level. VaR is widely used
in financial risk assessment.
o Formula: If X is the return and α\alphaα is the confidence
level, VaR is computed by: VaRα=−quantileα(X)
Summary
Risk analysis is essential for making informed decisions in business and
economics. By combining probability with risk assessment, managers
can evaluate the potential impact of different choices under conditions of
risk and uncertainty. Decision rules and valuation models offer
structured approaches to measure, analyze, and manage risks, ensuring
that managerial decisions are based on a comprehensive understanding
of potential outcomes.