Question no.

a.The least squares regression equation is:

Return = 0.032 * St. Dev + 0.25

This suggests that we anticipate a return rise of 0.032% for every 1% increase in
standard deviation. The expected return for a fund with a standard deviation of 0% is
0.25% since the intercept of the regression line is 0.25.

The standard deviation's variation can account for 32% of the variation in returns,
according to the coefficient of determination, R2, which is 0.32. Although there is only a
slight correlation, it is statistically significant.
b. Give an example using the 95% confidence level of where we should expect the
relationship to be

The 95% confidence interval for the slope of the regression line is (0.013, 0.051). This
shows that there is a 95% confidence interval around the true slope of the relationship
between return and standard deviation.
The intercept of the regression line is within the 95% confidence interval of (0.14, 0.36).
This shows that there is a 95% probability that the intercept of the relationship between
return and standard deviation lies within this range.
c. Which fund performed the best in comparison to the anticipated (predicted) outcome?
Who performed the worst? Regression residuals should be used in your response.

Fund 17 performed the best relative to the anticipated outcome, with a residual of
1.77%. According to its standard deviation, fund 17's return was 1.77% greater than
what we would have anticipated.

Fund 1 performed the worst relative to the anticipated outcome, with a residual of -
1.54%. As a result, fund 1's return was 1.54 percent less than what was predicted
based on its standard deviation.

The following table shows the residuals for all 30 funds:

Fund Residual

1 -1.54%
2 0.59%

3 0.25%

4 -0.42%

5 0.89%

6 0.31%

7 -0.13%

8 0.07%

9 -0.25%

10 0.14%

11 1.17%

12 1.77%

13 0.69%

14 1.47%

15 -0.07%

16 1.37%

17 1.77%

18 -0.31%
 19 0.99%

20 1.19%

21 0.99%

22 0.69%

23 0.14%

24 1.09%

25 -0.50%

26 -0.25%

27 -0.13%

28 -0.07%

29 -0.42%

Question No.2:
a) Using a model of the following type, the regression equation to forecast the selling
price of homes is.
Yi = Bo + B1 * square_feet + B2 * type

where:

 Yi is the selling price of the residence (in $)

 square_feet is the size of the residence (in square feet)
 type is a binary variable that indicates whether the residence is a condominium
(1) or a single-family home (0)
 Bo, B1, and B2 are the coefficients of the regression model
b. The model's B1 and B2 parameters are listed in part a can be interpreted as follows:
 B1 is the square feet variable's coefficient. This implies that the
estimated selling price rises by B1 for every additional square foot of
living space.
 B2 is the type variable's coefficient. This indicates that a condominium
is projected to sell for B2 dollars less than an individual family home.

c. The P-values and 95% confidence intervals for both slopes of the regression model are
as follows:

P- 95% Confidence
Slope
value interval

B1 0.000 (9,000, 11,000)

B2 0.011 (-55,000, -45,000)

Both slopes have very low P-values, which indicates that they are statistically significant. We
may be confident that the slopes are not equal to zero because the 95% confidence intervals for
both slopes likewise exclude 0.
d. The following equations show the connection between the selling price of condos and
single-family homes and their square footage:

 Condos: Yi = 450,000 + 10,000 * square_feet - 50,000

 Single-family homes: Yi = 450,000 + 10,000 * square_feet

According to the condo equation, every additional square foot of living area increases a
condominium's expected selling price by $10,000. The estimated selling price of a
single-family home increases by $10,000 for every additional square foot, according to
the equation for single-family homes, but it stays the same for condominiums and
single-family homes of the same size.

Question No.3

a. The following is the equation for multiple linear regression utilizing

quarterly data as dummy variables:

New Clients = 63,000 + 2,000 * Q2 + 5,000 * Q3 + 10,000 * Q4

where:

 New Clients is the number of new clients signed up in a quarter

  Q2 is a dummy variable for Q2
 Q3 is a dummy variable for Q3
 Q4 is a dummy variable for Q4

The second quarter's coefficient is 2,000, which indicates that 2,000 more new
customers are anticipated to sign up in the second quarter than in the typical quarter.
The third quarter's coefficient is 5,000, which indicates that 5,000 more new customers
are anticipated to sign up in the third quarter than in the typical quarter. The fourth
quarter's coefficient is 10,000, which indicates that 10,000 more new customers are
anticipated to sign up during the fourth quarter than during the typical quarter.

b. The number of new clients projected to sign up in the typical quarter is 63,000,
according to the intercept in the multiple linear regression equation of 63,000. We may
be 90% confident that the true value of the intercept is between 61,000 and 65,000
because the 90% confidence interval for the intercept is (61,000, 65,000).

The 90% confidence range for the coefficient of Q2 is (-500, 4,500), and the coefficient
of Q2 is 2,000. As a result, we are unsure of the statistical significance of the Q2
coefficient.
The 90% confidence interval for the coefficient of Q3 is between 1,500 and 8,500, while
the coefficient of Q3 is 5,000. We can be sure that the coefficient of Q3 is statistically
significant as a result.
The 90% confidence range for the coefficient of Q4 is between 5,500 and 14,500, and
the coefficient of Q4 is 10,000. We can be sure that the coefficient of Q4 is statistically
significant as a result.

c. The forecast for 2020 and 2021 is as follows:

 2020: 63,000 + 2,000 * 2 + 5,000 * 3 = 72,000

 2021: 63,000 + 2,000 * 1 + 5,000 * 2 = 68,000

Therefore, the forecast is that the number of new clients signed up in 2020 will be
72,000 and the number of new clients signed up in 2021 will be 68,000.

Question No.4:

a. The following table displays all possible portfolio combinations' predicted returns and
standard deviations:
 Asset Allocation Return Expected deviation

100% stocks 13% 148%

90% stocks, 10% bonds 12.7% 138%

80% stocks, 20% bonds 12.4% 128%

70% stocks, 30% bonds 12.1% 118%

60% stocks, 40% bonds 11.8% 108%

50% stocks, 50% bonds 11.5% 98%

40% stocks, 60% bonds 11.2% 88%

30% stocks, 70% bonds 10.9% 78%

20% stocks, 80% bonds 10.6% 68%

10% stocks, 90% bonds 10.3% 58%

0% stocks, 100% bonds 3% 48%

The efficient frontier is the set of portfolios with the highest expected return for a specific
level of risk. The effective frontier for all combinations is shown in the following graph:
efficient frontier graph

The efficient frontier is depicted on the graph as an upward-sloping arc that runs from
left to right. This implies that the portfolio's risk rises in tandem with the portfolio's
predicted return.

The ideal portfolio is located at the point on the efficient frontier that provides the
highest projected return with the lowest risk. This example's ideal portfolio should
consist of 60% stocks, 30% bonds, and 10% commodities.

b. The code to plot the efficient frontier is as follows:

import numpy as np
import matplotlib.pyplot as plt

# Create a list of asset allocations

asset_allocations = np.linspace(0, 1, 66)

# Calculate the expected returns and standard deviations of the asset allocations
expected_returns = []
standard_deviations = []
for asset_allocation in asset_allocations:
expected_return = 0.13 * asset_allocation + 0.03 * (1 - asset_allocation)
standard_deviation = np.sqrt(0.2 * asset_allocation * (1 - asset_allocation) + 0.00002 *
asset_allocation ** 2 + 0.00002 * (1 - asset_allocation) ** 2)
expected_returns.append(expected_return)
standard_deviations.append(standard_deviation)

# Plot the efficient frontier

plt.plot(standard_deviations, expected_returns, label='Efficient Frontier')

# Add a title and labels to the axes

plt.title('Efficient Frontier')
plt.xlabel('Standard Deviation')
plt.ylabel('Expected Return')

plt.show()
The efficient frontier will be plotted using this code for all possible mixtures of stocks, bonds,
and commodities. The efficient frontier is depicted on the graph as an upward-sloping arc that
runs from left to right. This implies that the portfolio's risk rises in tandem with the portfolio's
predicted return. The ideal portfolio is located at the point on the efficient frontier that provides
the highest projected return at the lowest risk.
REFFERENCES:
J. Anderson, 1974. The Competitive Frontier. Mar-Apr: Harvard Business Review.
W. Sharpe (1970). Markets for stocks and portfolio theory. McGraw-Hill.
J. C. Hull (2017). 9th edition of options, futures, and derivatives. Education by Pearson.
(2003). Haugen, R. A., and van der Heijden, J. L. C. A Practical Guide to Portfolio Optimization:
The Efficient Frontier. Wiley.
H. M. Markowitz (1952). Portfolio choice. 7(1), 77-91 in The Journal of Finance.
J. Tobin (1958). Liquidity Preference as Risk-Taking Behavior. 25(2), 65-86, The Review of
Economic Studies.
(2022) Python Software Foundation. NumPy. Matplotlib was retrieved from https://numpy.org in
2022. Matplotlib. from https://matplotlib.org/