Pica Daniel, Group 2

Homework 2

1. (a) What kinds of variables are likely to be non-stationary? How can such variables be
made stationary?
All economic variables are likely to be non-stationary. These types of variables can be made
stationary through differentiation of order 2.

(b) Why is it in general important to test for non-stationarity in time series data before
attempting to build an empirical model?
The stationarity or otherwise of a series can strongly influence its behavior and properties. Also,
due to spurious regressions. If two variables are trending over time, a regression of one on the
other could have a high R2 even if the two are totally unrelated

(c) Define the following terms and describe the processes that they represent
(i) Weak stationarity -> time series (+they are time independent)
(ii) Strict stationarity -> distribution function
(iii) Deterministic trend -> a+bt
(iv) Stochastic trend. -> yt = yt-1 + residual

2. A researcher wants to test the order of integration of some time series data. He decides to
use the DF test. He estimates a regression of the form „∆yt = μ + ψyt−1 + ut” and obtains
the estimate ψ = −0.02 with standard error = 0.31.
(a) What are the null and alternative hypotheses for this test?
H0 = ψ = 0
H1 = ψ < 0

(b) Given the data, and a critical value of −2.88, perform the test.
−0.02
= -0.064 => we accept H0 -> The series is non-stationary
0.31

(c) What is the conclusion from this test and what should be the next step?
We need to make the difference of order 1 and apply the Dickey-Fuller test for ∆yt.

(d) Why is it not valid to compare the estimated test statistic with the corresponding critical
value from a t-distribution, even though the test statistic takes the form of the usual t-ratio?
Because it is not student distributed (residuals are normal), but it is a Dickey-Fuller distribution
(residuals are not normal).

3. Using the same regression as for question 2, but on a different set of data, the researcher
now obtains the estimate ψ = −0.52 with standard error = 0.16.
(a) Perform the test.
−0.52
0.16
= -3.25
Pica Daniel, Group 2

(b) What is the conclusion, and what should be the next step?
The series is stationary, so, we accept H1.

(c) Another researcher suggests that there may be a problem with this methodology since it
assumes that the disturbances (ut ) are white noise. Suggest a possible source of difficulty
and how the researcher might in practice get around it.
We should use the augmented Dickey-Fuller method.

4. Consider a series of values for the spot and futures prices of a given commodity. In the
context of these series, explain the concept of cointegration. Discuss how a researcher might
test for cointegration between the variables using the Engle–Granger approach.

5. (a) Briefly outline Johansen’s methodology for testing for cointegration between a set of
variables in the context of a VAR.
The Johansen setup does permit the testing of hypotheses about the equilibrium relationships
between the variables. Johansen allows a researcher to test a hypothesis about one or more
coefficients in the cointegrating relationship by viewing the hypothesis as a restriction on the ∏
matrix. If there exist r cointegrating vectors, only these linear combinations or linear
transformations of them, or combinations of the cointegrating vectors, will be stationary.

(b) A researcher uses the Johansen procedure and obtains the following test statistics:
r λmax 95% critical value
0 38.962 33.178
1 29.148 27.169
2 16.304 20.278
3 8.861 14.036
4 1.994 3.962
Pica Daniel, Group 2

Determine the number of cointegrating vectors.

The number of cointegrating vectors is 2.

(c) ‘If two series are cointegrated, it is not possible to make inferences regarding the
cointegrating relationship using the Engle–Granger technique since the residuals from the
cointegrating regression are likely to be autocorrelated.’ How does Johansen circumvent
this problem to test hypotheses about the cointegrating relationship?
By considering all potential relationships through a vectorial approach.

(d) Give one or more examples from the academic finance literature of where the Johansen
systems technique has been employed.
The relationship between currencies.

6. (a) Suppose that a researcher has a set of three variables, yt (t = 1, . . . , T ), i.e. yt denotes
a p-variate, or p × 1 vector, that she wishes to test for the existence of cointegrating
relationships using the Johansen procedure. What is the implication of finding that the
rank of the appropriate matrix takes on a value of (i) 0 (ii) 1 (iii) 2 (iv) 3?
(i) no cointegrating vectors
(ii) one cointegrating vectors
(iii) two cointegrating vectors
(iv) three cointegrating vectors.

(b) The researcher obtains results for the Johansen test using the variables outlined in part
(a) as follows:
r λmax 5% critical value
0 38.65 30.26
1 26.91 23.84
2 10.67 17.72
3 8.55 10.71
Determine the number of cointegrating vectors.
The number of cointegrating vectors is 2.

7. Compare and contrast the Engle–Granger and Johansen methodologies for testing for
cointegration and modelling cointegrated systems. Which, in your view, represents the
superior approach and why?
Johansen methodology is the better one because it is able to determine the total number of
cointegrating vectors.

8. In EViews, open the ‘currencies.wf1’ file that will be discussed in detail in the following
chapter. Determine whether the exchange rate series (in their raw levels forms) are non-
stationary. If that is the case, test for cointegration between them using both the Engle–
Granger and Johansen approaches.
Pica Daniel, Group 2

Engle–Granger approach
Pica Daniel, Group 2

Johansen approach