Type author name/s here

Dougherty

Introduction to Econometrics,
5th edition
Chapter heading
Chapter 12: Autocorrelation

© Christopher Dougherty, 2016. All rights reserved.

 CONSEQUENCES OF AUTOCORRELATION

Yt   1   2 X t  ut
T
ˆ2   2   at ut
t 1

Xt  X
at  T

 sX  X 2

s 1

The consequences of autocorrelation for OLS are similar to those of heteroscedasticity. In

general, the regression coefficients remain unbiased, but OLS is inefficient because one
can find an alternative regression technique that yields estimators with smaller variances.
1
 CONSEQUENCES OF AUTOCORRELATION

Yt   1   2 X t  ut
T
ˆ2   2   at ut
t 1

Xt  X
at  T

 sX  X 2

s 1

The other main consequence is that autocorrelation causes the standard errors to be
estimated wrongly, often being biased downwards. Finally, although in general OLS
estimates are unbiased, there is an important special case where they are biased.
2
 CONSEQUENCES OF AUTOCORRELATION

Yt   1   2 X t  ut
T
ˆ2   2   at ut
t 1

Xt  X
at  T

 sX  X 2

s 1

Unbiasedness is easily demonstrated, provided that Assumption C.7 is satisfied. In the

case of the simple regression model shown, we have seen that the OLS estimator of b2 can
be decomposed as the second line where the at are as defined in the third line.
3
 CONSEQUENCES OF AUTOCORRELATION

Yt   1   2 X t  ut
T
ˆ2   2   at ut
t 1

Xt  X
at  T

 sX  X 2

s 1

 T

E   2    2  E   at ut 
ˆ
 t 1 
T T
  2   E  at ut    2   E  at  E  ut 
t 1 t 1

Now, if Assumption C.7 is satisfied, at and ut are distributed independently and we can write
the expectation of ̂ 2 as shown. At no point have we made any assumption concerning
whether ut is, or is not, subject to autocorrelation.
4
 CONSEQUENCES OF AUTOCORRELATION

Yt   1   2 X t  ut
T
ˆ2   2   at ut
t 1

Xt  X
at  T

 sX  X 2

s 1

 T

E   2    2  E   at ut 
ˆ
 t 1 
T T
  2   E  at ut    2   E  at  E  ut 
t 1 t 1

All that we now require is E(ut) = 0 and this is easily demonstrated.

5
 CONSEQUENCES OF AUTOCORRELATION

ut  ut 1   t

ut 1  ut  2   t 1

ut   2 ut  2   t 1   t

For example, in the case of AR(1) autocorrelation, lagging the process one time period, we
have the second line. Substituting for ut–1 in the first equation, we obtain the third.

6
 CONSEQUENCES OF AUTOCORRELATION

ut  ut 1   t

ut 1  ut  2   t 1

ut   2 ut  2   t 1   t

ut   t   t 1   2 t  2  ...

E ut   E  t   E  t 1    2 E  t  2   ...  0

Continuing to lag and substitute, we can express ut in terms of current and lagged values of
et with diminishing weights. Since, by definition, the expected value of each innovation is
zero, the expected value of ut is zero.
7
 CONSEQUENCES OF AUTOCORRELATION

ut  0 t  1 t 1  2 t 2  3 t 3

E ut   0 E  t   1 E  t 1    2 E  t  2    3 E  t  3   0

For higher order AR autocorrelation, the demonstration is essentially similar. For moving
average autocorrelation, the result is immediate.

8
 CONSEQUENCES OF AUTOCORRELATION

Yt   1   2 X t  ut
T
ˆ2   2   at ut
t 1

Xt  X
at  T

 sX  X 2

s 1

 T

E   2    2  E   at ut 
ˆ
 t 1 
T T
  2   E  at ut    2   E  at  E  ut 
t 1 t 1

For multiple regression analysis, the demonstration is the same, except that at is replaced
by at*, where at* depends on all of the observations on all of the explanatory variables in the
model.
9
 CONSEQUENCES OF AUTOCORRELATION

Yt   1   2 X t  ut
T
ˆ2   2   at ut
t 1

Xt  X
at  T

 sX  X 2

s 1

 T

E   2    2  E   at ut 
ˆ
 t 1 
T T
  2   E  at ut    2   E  at  E  ut 
t 1 t 1

We will not pursue analytically the other consequences of autocorrelation. An important

one is that the Gauss–Markov theorem, which guarantees the efficiency of the OLS
estimators, does not apply, since its proof requires no autocorrelation.
10
 CONSEQUENCES OF AUTOCORRELATION

Yt   1   2 X t  ut
T
ˆ2   2   at ut
t 1

Xt  X
at  T

 sX  X 2

s 1

 T

E   2    2  E   at ut 
ˆ
 t 1 
T T
  2   E  at ut    2   E  at  E  ut 
t 1 t 1

Another is that the expressions for the standard errors are invalid since they are based on
the assumption that there is no autocorrelation.

11
 CONSEQUENCES OF AUTOCORRELATION

Yt   1   2 X t   3Yt 1  ut

ut  ut 1   t

Now we come to the special case where OLS yields inconsistent estimators if the
disturbance term is subject to autocorrelation.

12
 CONSEQUENCES OF AUTOCORRELATION

Yt   1   2 X t   3Yt 1  ut

ut  ut 1   t

If the model specification includes a lagged dependent variable, OLS estimators are biased
and inconsistent if the disturbance term is subject to autocorrelation. This will be
demonstrated for AR(1) autocorrelation and an ADL(1,0) model with one X variable.
13
 CONSEQUENCES OF AUTOCORRELATION

Yt   1   2 X t   3Yt 1  ut

ut  ut 1   t

Yt 1   1   2 X t 1   3Yt  2  ut 1

Lagging the ADL(1,0) model by one time period, we obtain the third line. Thus Yt–1 depends
on ut–1. As a consequence of the AR(1) autocorrelation ut also depends on ut–1.

14
 CONSEQUENCES OF AUTOCORRELATION

Yt   1   2 X t   3Yt 1  ut

ut  ut 1   t

Yt 1   1   2 X t 1   3Yt  2  ut 1

Hence we have a violation of part (1) of Assumption C.7. The explanatory variables, Yt–1, is
not distributed independently of the disturbance term. As a consequence, OLS will yield
inconsistent estimates.
15
 CONSEQUENCES OF AUTOCORRELATION

Yt   1   2 X t   3Yt 1  ut

ut  ut 1   t

Yt 1   1   2 X t 1   3Yt  2  ut 1

This was described as a special case, but actually it is an important one. ADL models are
frequently used in time series regressions and autocorrelation is a common problem.

16
Copyright Christopher Dougherty 2016.

These slideshows may be downloaded by anyone, anywhere for personal use.

Subject to respect for copyright and, where appropriate, attribution, they may be
used as a resource for teaching an econometrics course. There is no need to
refer to the author.

The content of this slideshow comes from Section 12.1 of C. Dougherty,

Introduction to Econometrics, fifth edition 2016, Oxford University Press.
Additional (free) resources for both students and instructors may be
downloaded from the OUP Online Resource Centre
http://www.oxfordtextbooks.co.uk/orc/dougherty5e/.

Individuals studying econometrics on their own who feel that they might benefit
from participation in a formal course should consider the London School of
Economics summer school course
EC212 Introduction to Econometrics
http://www2.lse.ac.uk/study/summerSchools/summerSchool/Home.aspx
or the University of London International Programmes distance learning course
20 Elements of Econometrics
www.londoninternational.ac.uk/lse.

2016.05.22