GRA6515 Quantitative Methods for Finance

Practice Exercises: Hypothesis Testing

1. Define the following terms:

a. Null hypothesis.
b. Alternative hypothesis.
c. Test statistic.
d. Type I error.
e. Type II error.
f. Power of a test.
g. Rejection point (critical value).

2. Suppose that, on the basis of a sample, we want to test the hypothesis that the mean debt-
to-total-assets ratio of companies that become takeover targets is the same as the mean
debt-to-total-assets ratio of companies in the same industry that do not become takeover
targets. Explain under what conditions we would commit a Type I error and under what
conditions we would commit a Type II error.
If we rejected the hypothesis of the equality of the mean debt-to-total-assets ratios of takeover
target companies and non-takeover-target companies when, in fact, the means were equal, we
would be committing a Type I error. On the other hand, if we failed to reject the equality
of the mean debt-to-total-assets ratios of takeover target companies and non-takeover-target
companies when the means were different, we would be committing a Type II error.

3. Suppose we are testing a null hypothesis, H0 , versus an alternative hypothesis, H1 , and

the p-value for the test statistic is 0.031. At which of the following levels of significance —
α  0.10, α  0.05, and/or α  0.01 — would we reject the null hypothesis?

4. Identify the appropriate test statistic or statistics for conducting the following hypothesis
tests. (Clearly identify the test statistic and, if applicable, the number of degrees of freedom.
For example, “We conduct the test using an x-statistic with y degrees of freedom.”)

a. H0 : µ  0 versus H1 : µ  0, where µ is the mean of a normally distributed popula-

tion with unknown variance. The test is based on a sample of 15 observations. The
appropriate test statistic is a t-statistic with n
1  15
1  14 degrees of freedom.
b. H0 :µ  0 versus H1 : µ  0, where µ is the mean of a normally distributed popula-
tion with unknown variance. The test is based on a sample of 40 observations. The
appropriate test statistic is a t-statistic with 40 - 1 = 39 degrees of freedom.
c. H0 : µ ¤ 0 versus H1 : µ ¡ 0, where µ is the mean of a normally distributed population
with known variance σ 2 . The sample size is 45. The appropriate test statistic is a z-
statistic because the sample comes from a normally distributed population with known
variance.
d. H0 : σ 2  200 versus H1 : σ 2  200, where σ 2 is the variance of a normally distributed
population. The sample size is 50. The appropriate test statistic is chi-square (χ2 ) with
50 - 1 = 49 degrees of freedom.
e. H0 : σ12  σ22 versus H1 : σ12  σ22 , where σ12 is the variance of one normally distributed
population and σ22 is the variance of a second normally distributed population. The
test is based on two independent random samples. The appropriate test statistic is the
F-statistic (the ratio of the sample variances).

1
 f. H0 : (Population mean 1) - (Population mean 2) = 0 versus
H1 : (Population mean 1) - (Population mean 2)  0, where the samples are drawn
from normally distributed populations with unknown variances. The observations in
the two samples are correlated. The appropriate test statistic is a t-statistic for a paired
observations test (a paired comparisons test), because the samples are correlated.
g. H0 : (Population mean 1) - (Population mean 2) = 0 versus
H1 : (Population mean 1) - (Population mean 2)  0, where the samples are drawn
from normally distributed populations with unknown but assumed equal variances.
The observations in the two samples (of size 25 and 30, respectively) are independent.
The appropriate test statistic is a t-statistic using a pooled estimate of the population
variance. The t-statistic has 25 + 30 - 2 = 53 degrees of freedom.
5. Willco is a manufacturer in a mature cyclical industry. During the most recent industry
cycle, its net income averaged $30 million per year with a standard deviation of $10 million
(n  6 observations). Management claims that Willco’s performance during the most
recent cycle results from new approaches and that we can dismiss profitability expectations
based on its average or normalized earnings of $24 million per year in prior cycles.
a. With µ as the population value of mean annual net income, formulate null and al-
ternative hypotheses consistent with testing Willco management’s claim. H0 : µ ¤ 24
versus H1 : µ ¡ 24.
b. Assuming that Willco’s net income is at least approximately normally distributed,
identify the appropriate test statistic. Given that net income is normally distributed
with unknown variance, the appropriate test statistic is t with n
1  6
1  5 degrees
of freedom.
c. Identify the rejection point or points at the 0.05 level of significance for the hypothesis
tested in Part a. The rejection point (critical value) of 2.015. We will reject the null if
t ¡ 2.015.
d. Determine whether or not to reject the null hypothesis at the 0.05 significance level.
We do not reject the null hypothesis.
6. Altman and Kishore (1996), in the course of a study on the recovery rates on defaulted
bonds, investigated the recovery of utility bonds versus other bonds, stratified by seniority.
The following table excerpts their findings.

Recovery Rates by Seniority

Industry Group Ex-Utilities Sample
Industry Group/ Number of Average Standard Number of Average Standard
Seniority Observations Price Deviation Observations Price Deviation
Public Utilities
Senior Unsecured 32 $77.74 $18.06 189 $42.56 $24.89

Assume that the populations are normally distributed and that the samples are indepen-
dent. The population variances are unknown; do not assume they are equal. The test
hypotheses are H0 : µ1
µ2  0 versus Ha : µ1
µ2  0, where µ1 is the population mean
recovery rate for utilities and µ2 is the population mean recovery rate for non-utilities.

a. Calculate the test statistic.

t-statistic for difference in means, unknown and unequal variance.
X1
X2
TS a  9.585
s21 {n1 s22 {n2

2
 b. Determine whether to reject the null hypothesis at the 0.01 significance level without
reference to degrees of freedom.
The value of the test statistic is quite large. With degrees of freedom highly likely being
larger than 3, the null hypothesis of equal population means is rejected.
c. Calculate the degrees of freedom.
55 degrees of freedom:

 ppss1{{nnq1 s22 {n2 q2

2
df 2
1 1
2 ps22 {n2 q2
 54.93  55
n1 n2

d. How would the solution of the problem change if you assumed that the variances are
equal?
t-statistic for difference in means, unknown and unequal variance. The number of degrees
of freedom is df  n1 n2
2  219.

 pn1
n1qs1 n pn
22
1qs2  577.98

2 2
s2p
1 2
X1
X2
TS  b  7.655
s2p {n1 s2p {n2

The test statistic is going to be still larger than the critical value Ñ reject the null
hypothesis.

7. You are investigating whether the population variance of returns on the S&P 500/BARRA
Growth Index changed subsequent to the October 1987 market crash. You gather the
following data for 120 months of returns before October 1987 and for 120 months of
returns after October 1987. You have specified a 0.05 level of significance.

Mean Monthly
Time Period n Return (%) Variance of Returns
Before October 1987 120 1.416 22.367
After October 1987 120 1.436 15.795

a. Formulate null and alternative hypotheses consistent with the verbal description of
the research goal.
ore  σaf ter vs. Ha : σbef ore  σaf ter
2
H0 : σbef 2 2 2

b. Identify the test statistic for conducting a test of the hypotheses in Part a.
The test statistic is T S  s21 {s22 . Under the null hypothesis the test statistic is distributed
according to the Fisher-distribution with 119 and 119 degrees of freedom.
c. Determine whether or not to reject the null hypothesis at the 0.05 level of significance.
The test is two-sided. First, determine the critical value CV  F119,119 p0.05{2q  1.435.
Next, compute the test statistic: T S  22.367{15.795  1.416. Since T S CV , we fail
to reject the null hypothesis that the return variance remained the same after October
1987.
Note: Generally the test of variances has a lower and an upper critical values. The lower
bound in our example is CV   F119,119 p1
0.05{2q  0.6969. Whether you compute
the test statistic as ratio of the larger variance over the smaller variance (1.416) or the
other way around (0.706), the null hypothesis is rejected when T S ¡ CV or T S CV  .

3
8. During a 10-year period, the standard deviation of annual returns on a portfolio you are
analyzing was 15 percent a year. You want to see whether this record is sufficient evidence
to support the conclusion that the portfolio’s underlying variance of return was less than
400, the return variance of the portfolio’s benchmark.

(a) Formulate null and alternative hypotheses consistent with the verbal description of
your objective. We have a “less than” alternative hypothesis, where σ 2 is the variance
of return on the portfolio. The hypotheses are H0 : σ 2 ¥ 400 versus H1 : σ 2 400, where
400 is the hypothesized value of variance, σ02 .
(b) Identify the test statistic for conducting a test of the hypotheses in Part a. The test
statistic is chi-square with 10
1  9 degrees of freedom.
(c) Identify the rejection point or points at the 0.05 significance level for the hypothesis
tested in Part a. 3.325
(d) Determine whether the null hypothesis is rejected or not rejected at the 0.05 level of
significance. We do not reject the null hypothesis.