Sie sind auf Seite 1von 14

1

QUANTITATIVE METHODS II
Quiz 1
Saturday, October 23, 2010

Time : 120 minutes

Total No. of Pages : 14 Name ________________________

Total No. of Questions: 3 Roll No. ________________________

Total marks: 50 Section: _______________________

Instructions

1. This is a Closed Book Exam. You are not allowed to carry anything other than stationary and calculator.
2. Answer all questions only in the space provided following the question.
3. Show all work and give adequate explanations to get credit.
4. You may use the backside of the last page for rough work only if needed. Do NOT attach any rough
work/sheets.
5. Encircle or underline your final answer for each part.
6. No clarifications will be made during the exam.
2

Question 1 (30 points)

Politicians believe that economic condition is the greatest concern of voters. Therefore, the percentage
of votes (PV) the incumbent gets depends on macroeconomic variables such as the inflation rate (IR),
the unemployment rate (UR) and the growth rate of disposable income (DI). The following table
contains the relevant data.

Table 1: PV, IR, UR and DI

PV DI UR IR UR+IR
45 1.2 8.3 3.5 11.8
48 1.8 7.4 3.8 11.2
49.5 2 7.1 3.9 11
48.8 1.9 6.5 4.2 10.7
50.4 2.2 6.2 4.6 10.8
51.3 2.4 5.9 4.9 10.8
52.4 2.7 5.7 5 10.7
47.6 1.9 7 3.4 10.4
54.1 3 5.1 5.1 10.2
50 2.3 6.1 3.4 9.5

Where:

PV = Percentage of votes the incumbent gets

DI = Rate of increase in disposable Income

UR = Unemployment rate

IR = Inflation Rate

Mean values of variables are:


PV 49.71, DI 2.14, UR 6.53, IR 4.18

Sum of squared deviations from mean values:


( PV PV ) 2 60.029 , ( DI DI ) 2 2.284 , (UR UR) 2 7.861 ,

( IR IR) 2 4.116
The following model is used to estimate the PV and the regression output is shown in Table 2.
3

Model 1: PV 0 1DI (1)

Table 2

SUMMARY OUTPUT Formatted Table


Regression Statistics
Multiple R 0.986059
R Square 0.972313
Adjusted R
Square 0.972313
Standard Error 0.455803
Observations 10

ANOVA
Df SS MS F Significance F
Regression 1 58.366952 58.366952 280.9399 7.57 Formatted: Right: 0.31"
Residual 8 1.662048 0.207756
Total 9 60.029
Standard Upper
Coefficients Error t Stat P-value Lower 95% 95%
Intercept 38.89194 0.661319 58.80964 7.76E-12 37.36694 40.41695
DI 5.055166 0.301598 16.76127 4.359679 5.750653

Answer the following questions using table 2.

Question 1.1

Calculate the change in the average percentage of votes the incumbent gets (PV) when the rate of
disposable income increases by one unit? (3 points)

5.055166

Question 1.2

At 95% confidence level, calculate the maximum value of PV when the value of DI 2.3. (3 points)

50.5188218 + 2.306*0.455803*sqrt(1+.1+.011208406) = 51.62680772


4

Question 1.3

At 95% confidence level, test the hypothesis that incumbents with DI of 2.3 on average get 52 percent of
the votes. (4 points)

Mean = 50.5188218

Max = 50.5188218 + 2.306*0.455803*sqrt(.1+.011208406) = 50.86933574

No

Question 1.4

The incumbent wins when she/he gets more than 50% votes. Check whether the probability that the
incumbent will win is greater than 99% when the value of DI is 2.5. (3 points)

Mean = 51.529865 - 2.896*0.455803*sqrt(1+.1+.011208406) = 50.13839613

Yes

Use Model 2 and the corresponding regression output to answer questions 1.5 - 1.7

Model 2:

PV 0 1DI 2UR 3 IR (2)

Regression output for Model 2


5

Regression Statistics
Multiple R 0.99329
R Square 0.986625
Adjusted R
Square 0.979939
Standard Error 0.365798
Observations 10

ANOVA
Significance
Df SS MS F F
Regression 3 59.226152 19.742050 147.540 9.78
Residual 6 0.802848 0.133808
Total 9 60.029

Standard Upper
Coefficients Error t Stat P-value Lower 95% 95%
Intercept 37.21994 5.405421 6.885669 23.99335 50.44653
DI 4.38335 0.944323 4.641791 2.072676 6.694024
UR 0.02555 0.508229 0.050273 -1.21804 1.269142
IR 0.70403 0.280186 2.512723 0.018438 1.389621

Question 1.5

State whether the overall model in equation (2) can be accepted at 99% confidence level using an
appropriate test. (4 points)

Yes

Question 1.6

Identify the explanatory variables in equation 2 that are significant at 95% confidence level. (3 points)

DI & IR

t-stat > t-crit for these


6

Question 1.7

For politicians, which explanatory variable is the most important variable? Justify your answer. (4
points)

Standardized beta: DI = 0.855, UR =, IR = 0.184

When Jimmy Carter was a candidate for the presidency, he coined a new term, the Misery Index (MI),
where, MI = UR + IR. Model 3 used to predict PV is shown in equation 3.

Model 3: PV 0 1DI 2 MI (3)

Regression output for model 3:

SUMMARY OUTPUT Formatted Table

Regression Statistics
Multiple R 0.991534155
R Square 0.983139982
Adjusted R
Square 0.978322834
Standard Error 2.629896196
Observations 10

ANOVA
Significance
Df SS MS F F
Regression 2 59.01691 29.50845 204.09067
Residual 7 1.012094 0.144585
Total 9 60.029
7

Question 1.8

Check whether the misery index is significant at 95% confidence level. (3 points)

Partial F = 4.495307259

F-crit = 5.59

Question 2 (15 points)

Highly publicized salaries of corporate chief executive officers (CEOs) in the United States have
generated sustained interest in understanding factors related to CEO compensation in general.
Data on the annual compensation of the CEOs of 167 financial companies is culled out from a
larger dataset that appeared in Forbes magazine. Each year, Forbes magazine publishes data
giving the compensation package of 800 top CEOs in the US including those of financial
companies.

The variable of interest is total compensation, defined as the sum of salary plus any bonuses,
including stock options. How does one explain the wide variation in total compensation of CEOs
in the same industry? Some of the variables considered that might help explain the variation in
salary are:

MBA?: is equal to 1 if the CEO has an MBA, 0 otherwise


MasterPhD?: is equal to 1 if the CEO has a masters or PhD degree, 0 otherwise.
YearsFirm(Yrs): Total number of years the CEO worked for the company for which he/she is
currently CEO
YearsCEO(Yrs): Total number of years the CEO has served the company as a CEO
StockOwned(%): % of company stock owned by the CEO
Sales(millions of $): Annual sales of the company in millions of dollars
Profits(millions of $): Annual profits of the company in millions of dollars

A regression model is built using SPSS with the Forward method giving it a choice of choosing
any of the above listed variables. The output obtained at the first step is as follows:

Table 1: Model Summaryb

Model Adjusted R Std. Error of Durbin-


R R Square Square the Estimate Watson
8

d
1 0.347 0.120 .115 4406976.255 2.031
i

a. Predictors: (Constant), Profits(in 000000 of $)


b. Dependent Variable: TotalComp($)

Table 2: Coefficientsa

Model Standardiz
Unstandardized ed
Coefficients Coefficients

Partial
Correlat
B Std. Error Beta t Sig. Correlation ion

1 (Constant) 1182515.538 404938.70 2.920 .004


7

Profits(in 4090.826 859.264 .347 4.761 .000 .347 .347


000000 of $)

a. Dependent Variable: TotalComp($)

Table 3: Excluded Variablesb

Model Collinearity

Partial Statistics

Beta In t Sig. Correlation Tolerance

1 MBA? -.016a -.018 .999


a
MasterPhD? -.065 -.069 .992

YearsFirm(Yrs) -.087a -.092 .989

YearsCEO(Yrs) .024a .026 .982


a
StockOwned(%) .021 .022 .981

Sales(millions of $) -.157a -.070 .174

a. Predictors in the Model: (Constant), Profits(in 000000 of $)


b. Dependent Variable: TotalComp($)
9

Using the information given above answer the followng questions:


a) The correlation and partial correlations given for Profit in Table 2 are the same. Explain
why. (1 point)

This is because there is only 1 variable that has been considered; there is no other value
whose effect has to be removed.

b) What is the proportion of explained variation compared to the total variation for the
model above? (2 points)

0.120

c) Which variable should be chosen as the next candidate to enter the regression equation?
Why? (1 point)

Years firm, highest partial correlation

d) What does the collinearlity statistics for Sales(millions of $) given in Table 3 in the
output mean? State clearly the steps needed to calculate it. (2 points)

Sales is correlated (1-.115) = 0.885 to the other X's


Regress sales w.r.t. others
10

Model Standardiz
Unstandardized ed
Coefficients Coefficients Correlations

Zero-
B Std. Error Beta T Sig. order Partial Part

1 (Constant) 1248165.7 246751.0 5.058 .000


45 35

MBA? 322221.97 .036 .066 .060


8

Sales(millions 256.372 45.960 .400 5.578 .000 .396 .399 .399


of $)
a. Dependent Variable: TotalComp($)

e) Are successive observations in the data used for the regression independent or not?
Explain? (1 point)

The variable MBA is added to the regression equation and the following information is obtained:

f) What is the impact on the compensation of CEOs who do not have an MBA? (1 point)

-322221.978

The information on outlier statistics obtained from SPSS is given below for the first 25
observations (out of the 167 observations).

g) Identify any CEO(s) whose compensation is an influential outlier. If you find any, is it
because the sales for this company are very high compared to all others? Explain.
(4 points)
11

Mahalanobis: > 11 (n>=30), > 15 (n>=100), > 25 (n>=500) Formatted: Indent: Left: 0"
Cook: > 4/(n-k-1)
Leverage: > 2(k+1)/n
SDFBeta: > 2/sqrt(n)
DFFit: > 2*sqrt(k+1/n)

h) Find the compensation predicted by the regression equation if the 9th company is not included in
the regression. The DFB0, DFB1 and DFB2 values given by SPSS for this company are :
76642.07, -75727.37 and -3.82 (3 points)

B0-dash = B0 - DFB0, ......


12

Where MAH = Mahalanobis distance, COOK = Cooks distance, LEV = Leverage values, SDF = Standardized
DFFit, SDBi = Standardized DFBeta values

Question 3 (8 points)

Company TotalComp MBA? Sales MAH COOK LEV SDF SDB0 SDB1 SDB2
($) (millions)
1 5095700 1 9176 5.4049 0.0035 0.0326 0.1026 -0.033 0.0613 0.0752
2 3409250 0 628 0.7011 0.0026 0.0042 0.0876 0.0871 -0.049 -0.03
3 1437240 0 1262 0.5831 1E-05 0.0035 -0.006 -0.006 0.0033 0.0013
4 8125370 0 16588 13.541 0.0445 0.0816 0.3659 -0.049 -0.047 0.3466 Formatted: Font: Bold
5 1356070 0 8799 3.1606 0.0075 0.019 -0.15 -0.027 0.0447 -0.12
6 3214100 1 1097 2.0106 0.0021 0.0121 0.08 0.0041 0.0641 -0.009
7 15914900 0 3396 0.5675 0.1121 0.0034 0.653 0.5129 -0.361 0.1343 Formatted: Font: Bold
8 1797910 0 3163 0.5406 4E-05 0.0033 -0.011 -0.009 0.0061 -0.002
9 10149200 0 4005 0.6711 0.039 0.004 0.3537 0.2501 -0.187 0.1129
10 1173000 0 429 0.7489 2E-05 0.0045 -0.008 -0.008 0.0046 0.0031
11 816667 0 1393 0.5652 0.0004 0.0034 -0.033 -0.032 0.0191 0.0067
12 2375840 0 1958 0.5133 0.0002 0.0031 0.0258 0.024 -0.015 -0.002
13 780250 0 328 0.7751 0.0002 0.0047 -0.025 -0.025 0.0137 0.0097
14 4485530 0 14173 9.4837 0.0007 0.0571 -0.045 0.0033 0.0074 -0.042 Formatted: Font: Bold
15 3364990 0 1078 0.612 0.0021 0.0037 0.0784 0.0769 -0.045 -0.021
16 1030600 0 1087 0.6105 0.0002 0.0037 -0.021 -0.021 0.0121 0.0055
17 2598470 1 4455 2.3966 2E-05 0.0144 -0.007 0.0011 -0.006 -0.003
18 5103330 1 5822 2.9713 0.0064 0.0179 0.1386 -0.031 0.1017 0.0697
19 2224530 0 5584 1.1629 0.0002 0.007 -0.023 -0.011 0.0101 -0.012
20 4576230 0 16054 12.579 0.0037 0.0758 -0.105 0.0128 0.0142 -0.099 Formatted: Font: Bold
21 697279 0 404 0.7553 0.0003 0.0046 -0.029 -0.029 0.0162 0.0111 Formatted: Font: Bold
22 16171700 0 251 0.796 0.1495 0.0048 0.7715 0.7708 -0.426 -0.313 Formatted: Font: Bold
23 663111 0 712 0.6825 0.0004 0.0041 -0.033 -0.033 0.0189 0.011
24 1177820 0 2617 0.505 0.0003 0.003 -0.03 -0.027 0.0175 -0.001
25 3659940 0 386 0.7599 0.0036 0.0046 0.1032 0.1029 -0.057 -0.04
A cab company has supplied the accompanying data, which show the number of accidents involving its
cabs over the past 2 years.
13

Year Winter Spring Summer Fall


2008 22 11 7 10
2009 25 12 9 12

Question 3.1

Use ratio-to-moving average to calculate the seasonality index of the four seasons described in the data
(4 points)

Question 3.2

Forecast the number of accidents for year 2010 winter (use method of moving averages to estimate the
trend component). (2 points)
14

Question 3.3

If single exponential smoothing is used to forecast the future value for the data given in this question,
what should be the value of the smoothing constant (high or low)? Justify your answer (2 points)

Das könnte Ihnen auch gefallen