Professional Documents
Culture Documents
Stat Assignment 2
Stat Assignment 2
Since n = 36 > 30 from unknown population distribution, by the Central Limit Theorem,
X
the
sampling
distribution
of
is unknown, so we use the t-distribution.
is
assume
approximately
normal.
X
76.8668.75
0
=
= 5.016
s / n
9.70 / 36
Reject H0 at 0.05 . There is sufficient evidence that the true mean monthly adjusted
closing price of the stock exceed its monthly adjusted closing price in Oct2011.
(b) The proportion of the past 36 months where the monthly adjusted closing price is
higher than the opening price must exceed 0.7.
Let be the population proportion of the past 36 months where the monthly adjusted
closing price is higher than the opening price must exceed 0.7.
H0: 0.7
H1: 0.7
Assume population follows binomial distribution. As n =36> 30, n 0 =25.2 > 5, n(1 0 )
10.8 5 so Normal approximation can be used, and the distribution of sample
p=
19
36
Z
= 0.528
p 0
0.528 0.7
2.252
0 (1 0 )
0.7(1 0.7)
n
36
Do not reject H0 at 0.05 . There is insufficient evidence to conclude that the proportion
of the past 36 months where the monthly adjusted closing price is higher than the opening
price must exceed 0.7.
To conclude, as HSBC Holdings plc does not fulfill both of the criteria, which indicates that
it is not a worth buying stock.
2. Develop a simple linear regression model to forecast the monthly adjusted closing price.
From the Excel output, b 0 =65.723 and b 1 =+0.602, the estimated simple linear regression equation is
Y^ adj close=65.723+0.602 X Month
where Y^
is the predicted monthly adjusted closing price for HSBC Holdings plc and X
b 1 = +0.602
For each increase in number of months, the expected monthly adjusted closing price for
HSBC Holdings is estimated to increase by 0.602 on average, keeping other things constant
r = +0.729
There is a moderate positive relationship between number of months and the monthly
adjusted closing price for HSBC Holdings.
2
r = 0.532
About 53.2% of total variation in monthly adjusted closing price for HSBC Holdings can
2
be explained by the regression model. However, r is not very large, showing that the model
might not be good to provide predication of monthly adjusted closing price or explain by
the number of months. Other factors should be identified so as to develop a better model
for future prediction.
80
70
60
Adjusted
Closing Price
50
40
30
20
10
0
8 12 16 20 24 28 32 36 40
Months
The regression model can only predict well if the new observation has an independent
variable value which is within the relevant range. If the independent variable value of the
new observation is out of the relevant range, we may still use the regression model to find
out the predicted value, but the accuracy of the predicted value will be doubted.
In daily life, stock price fluctuates, as there are many factors influencing it, including internal factors
such as company restructure and company reputations, as well as the external factors like inflation
and the number of substitutes available on the market. However, simple linear regression only
concerns about one variable (X). This clearly shows that simple linear regression would not be
comprehensive enough to predict the stock price. Although Hypothesis Testing helps to gather more
decisive information about characteristics of the entire population, but still there are probability to
reject a true null hypothesis or fail to reject a false null hypothesis.
In view of this, our group believes that multiple regression analysis would be a more reliable and
convincing way to estimate the stock price as it can make use of two or more variables, therefore we
can have a more accurate estimation by concerning more factors in the calculation. Moreover, the
regression model and hypothesis testing are just a reference for making financial decisions, we
should consider all dependencies criticality and realize that there are risks when investing in the
stock market.