Co-Integration, Error and Purchasing Power Parity between Canada and the

United States

David R. Johnson

The Canadian Journal of Economics / Revue canadienne d'Economique, Vol. 23, No. 4. (Nov.,
1990), pp. 839-855.

Stable URL:
http://links.jstor.org/sici?sici=0008-4085%28199011%2923%3A4%3C839%3ACEAPPP%3E2.0.CO%3B2-F

The Canadian Journal of Economics / Revue canadienne d'Economique is currently published by Canadian Economics
Association.

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained
prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in
the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/journals/cea.html.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic
journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers,
and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take
advantage of advances in technology. For more information regarding JSTOR, please contact [email protected].

http://www.jstor.org
Sat Mar 15 14:49:24 2008
Co-integration, error correction, and
purchasing power parity between
Canada and the United States
D A V I D R . J 0H N S0N Wilfrid Laurier University

Abstract. The concept of co-integration and the associated error-correction representation of

a dynamic process (see Engle and Granger 1987) is applied to the purchasing power parity
relationship between Canada and the United States. The results are supportive of purchasing
power parity (PPP)as a long-run equilibrium relationship between Canadian prices, American
prices, and the Canadian dollar / American dollar exchange rate. The PPP relationship is
maintained, although with different dynamic mechanisms, during both fixed and flexible
exchange rate regimes.

Cointe'gration, correction d'erreurs et parite' du pouvoir d'achat entre le Canada et les

Etats-Unis. Le concept de cointkgration et la representation qui y est associCe d'un processus
dynamique qui corrige ses erreurs (voir Engle & Granger 1987) sont utilises pour etudier
la relation de parite du pouvoir d'achat entre le Canada et les Etats-Unis. Les rCsultats
supportent la parite du pouvoir d'achat en tant que relation a long terme entre prix canadiens,
prix amCricains et taux de change entre les devises canadienne et amkricaine. La relation est
maintenue m&mesi differents mCcanismes dynamiques ont opere tant au cours de la periode
des changes fixes qu'au cours de la periode des changes flexibles.

I INTRODUCTION

Purchasing power parity, (PPP),is one of the foundations of price and exchange
rate behaviour in an open economy. Surveys of PPP are found in Officer (1976),
Katseli-Papaefstratiou (1979), and Levich (1985). Although PPP is firmly embedded
in virtually all discussions of exchange rate theory and policy, the validity of PPP
remains open to question. This study utilizes the error-correction mechanism and
the concept of co-integration associated with Engle and Granger (1987) to explore
the PPP relationship between Canada and the United States.

I would like to thank an anonymous referee for comments and the participants at meetings of the
Canadian Economics Association and the Summer Institute of the National Bureau of Economic
Research.

C a n a d ~ a nJournal of E c o n o m ~ c s Rebue canad~enned'Economlqoe, xxlil, No. 4

November novembre 1990 Prlntcd ~n Canada l m p r ~ m dail Canada

0008-4085 / 90 / 839-855 $1 .SO Canadian Economics Association

840 David R. Johnson

A major finding of this study is substantial evidence in favour of a PPP relation-

ship between Canadian and American prices and the Canadian dollar - American
dollar exchange rate, particularly over the 1950-86 period. There is evidence of
co-integration among the appropriate vectors of time series. The error-correction
terms (the deviations from the PPP relationship) are consistently significant in ex-
plaining changes in the exchange rate in periods of flexible exchange rates and are
usually significant in explaining changes in domestic prices in periods of either
flexible or fixed exchange rates.
The second major finding of this study is that different error-correction models
of exchange rate change and price change must be estimated from subperiods of the
data that correspond to different exchange rate regimes. If exchange rates are fixed,
adjustment towards PPP takes place only through the adjustment of the domestic
inflation rate. If exchange rates are flexible, then both the domestic inflation rate
and the level of the exchange rate can adjust in maintenance of PPP. The choice of
nominal exchange rate regime has a strong effect on the mechanism through which
PPP is maintained.
The paper proceeds as follows. The relationship of the present study to the
existing literature is considered in section 11. Section 111 outlines the implementation
of the tests for co-integration and the estimation of the error-correction models
including the PPP relationship. The results are presented and discussed in Section
IV.

11. R E L A T E D L I T E R A T U R E

The economic framework of Engle and Granger (1987) has two advantages over
previous approaches to the study of PPP: (I) The concept of co-integration deals
with the analysis of long-run equilibrium relationships between non-stationary time
series. Prices and exchange rates are non-stationary time series; PPP is their long-run
relationship. (2) In estimation of an error-correction model of price or exchange
rate change, a flexible short-run dynamic process is constrained to achieve long-
run equilibrium. Given these two important advantages, other researchers have
considered the validity of PPP in the framework of co-integrated variables.
Corbae and Ouliaris (1988), using a frequency-domain approach, use post-1973
data from six countries in a test of PPP. They find support for PPP only in the
Canadian-u s case. Mark (1987) uses the Engle-Granger time-domain approach to
test PPP after 1973 between the United States and five countries (including Canada).
He is unable to find support for PPP in the Canadian-American case. Baillie and
Selover (1987) and Taylor (1988), also with post-1973 data, cannot find support
for PPP in the Canadian-American case. Enders (1988) considers PPP between the
United States and three countries. He finds mixed evidence in favour of PPP in the
Canadian-American case.
The present paper differs from other studies of PPP in the co-integration frame-
work in two important ways. First, a long, in absolute time, series of observations is
used; thirty-six years of quarterly data and 100 years of annual data. There is some
 Co-integration, error correction 84 1

evidence that the length of sample is important in considering the properties of the
real exchange rate and the maintenance of PPP.Frankel (1986) finds support for the
idea that PPP between the United States and the United Kingdom increases as the
sample length increases. Hendry (1986) argues that the concept of co-integration
and the search for long-run economic relationships require not just large numbers
of observations but data over a long period of time. All previous studies of PPP in
the co-integration framework, except Enders (1988), consider only the post-1973
floating exchange rate period. However, Enders chooses to analyse only subsam-
ples (corresponding approximately to periods of fixed and flexible exchange rates)
drawn from the post-1960 period.
Second, this study estimates error-correction mechanisms for exchange rate
change and domestic price change that vary by choice of exchange rate regime.
The concept and measurement of PPP is not restricted to a specific exchange rate
regime; the mechanism for the maintenance of PPP is specific to each exchange rate
regime. When nominal exchange rates are flexible, both the nominal exchange rate
and the domestic inflation rate can respond to deviations from PPP.When nomi-
nal exchange rates are fixed, only the domestic inflation rate can adjust to correct
deviations from PPP.
Edison (1987) estimates an error-correction model of the u.s.-U.K.nominal ex-
change rate using annual data from 1890 to 1986. Part of the present study is the
estimation of an error-correction model of the u.s.-Canadian nominal exchange rate
using annual data from 1914 to 1986. Both the present study and Edison find some
evidence in favour of PPP.However, Edison does not estimate error-correction mod-
els that vary by exchange rate regime; or pre-test for co-integration of the relevant
variables; or estimate the associated error-correction model of domestic inflation.

111. C O I N T E G R A T I O N , E R R O R - C O R R E C T I O N , A N D P U R C H A S I N G P O W E R
PARITY

The PPP relationship is

P = KEP*
where the variables are
P = domestic currency (Canadian dollar) price index
P* = foreign currency (American dollar) price index
E = the exchange rate, the domestic currency (Canadian dollar) price of a unit
of foreign (American dollar) currency
K = a constant.
The operational definition of PPP,'relative PPP,' is that (1) holds in the long run for
a constant value of K.' The application of the concept of co-integration to the PPP
relationship is considered below.
1 K equals unity if there is perfect arbitrage between all goods in both countries and the bundles
that detine the national pnce indexes are composed of identical goods. This is unlikely to be the
case.
842 David R. Johnson

Co-integration
Engle and Granger (1987) define a co-integrated vector time series X t , denoted
c ~ ( db)
, if
1. each element of X, is integrated of order d
2. there exist up to r linear combinations of X, such that each element of a vector
times series Zt is integrated of order (d - b) where Z, is defined by

Co-integrated vectors occur frequently in macro-economic theory and there is a

burgeoning literature in their applications."
For the study of PPP the vector time series Xi has two elements and may be
written as

Previous studies have found that the logarithms of P,, P; and E, (when nominal
exchange rates are flexible) are integrated of order The logarithms of P,, PI",
E,, E,P: and P,IPf are in fact series with single unit roots over the sample in this
study.4
The normalized vector a is defined as

Co-integration requires that a when multiplied by either (3) or (4) yields a variable

2 Some exarnples are Engle and Granger (1987) and Campbell (1987) who study the consumption-
income relationship; Stock and Watson (1986) and Campbell and Shiller (1987) who study the
term structure of interest rates; and Hall (1986) and Jenkinson (1986) who study the behaviour of
real wages.
3 Wasserfallen (1986) and Nelson and Plosser (1982) cannot reject the hypothesis that many macro-
economic time series, including price indexes, have unit roots. Meese and Singleton (1982) show
that nominal exchange rates contain a unit root over the floating exchange rate period. The other
studies of PPP and co-integration as cited above also find that the variables in X, (as defined in
(4)) have unit roots.
4 In general the sum or difference of two series that individually contain a unit root also contains
a unit root. Co-integration is of interest precisely because the linear combination of integrated
variables is stationcuy. The standard tests for the presence of a unit root. a unit root with drift, or
a unit root with trend are found in Fuller (1976) and Dickey and Fuller (198 I ) .
 Co-integration, error correction 843

where Z, is stationary (integrated of order d - b = 0). A stationary variable has

finite variance, a constant mean, and a tendency to return to that mean in a finite
length of time. If Z, is stationary, then PPP is a meaningful concept. The difference
between Z, and its mean is interpreted as the percentage deviation of the domestic
prive level or the exchange rate from the PPP equilibrium level. The null hypothesis
is that the bivariate time series X,, (3) or (4), is not co-integrated. If this null
hypothesis is rejected, then the PPP relationship receives empirical ~ u p p o r t . ~

Error-correction models
Engle and Granger (1987) prove that if X, is co-integrated, then the change in X, has
an error-correction representation. This representation, also suggested by Davidson
et al. (1978), allows for much flexibility in short-run dynamics while the model is
constrained to return to long-run equilibrium.
The general error-correction representation is

where B is the backshift operator. Granger's Representation Theorem (1987) in-.

cludes the following results:
1. A*(O) ==I.
2. (1 - B)X, is a stationary stochastic process.
3. There exists some lag length in A*(B) where d(B) is unity.
4. t, is multivariate white noise.
5. Z,-l is defined as deviations from the cointegrating regression.
Expression (7) is the error-correction mechanism for any cointegrated vector time
series X,. The particular set of error-correction models estimated and their parameter
values depend on whether the nominal exchange rate regime is fixed or flexible.
The domestic price level can change in a fixed or flexible exchange rate regime.
The error-correction mechanism for the change in the domestic price level, the
inflation rate, using (7) and (3) is

5 There is also a substantial literature that tests the null hypothesis that the logarithm of the real
exchange rate contains a unit root (is non-stationary). The evidence is mixed. See Adler and
Lehman (1983), Roll (1973), Frenkel (l981), Mishkin (1984), Cumby and Obstfelt (1984),
Frankel (1986), and Huiringa (1987). Footnote 10 contains some results of unit root tests on
the data in this study. Testing for co-integration of the vector time series (3) or (4) is equivalent
to testing the null hypothesis that the logarithm of the real exchange rate is stationary. Since the
residuals from the co-integrating regression are important inputs into the error-correction model
this study tests for co-integration.
844 David R. Johnson

no and nl are sufficient to reduce the appropriate element o f the error term to
white noise. The term in parentheses following 71 is the lagged residual from the
co-integrating regression

In P, = cue + a1 In (E,PT)+ e,. (9)

The lagged residual from this regression, e,-l, is the error-correction term, the
estimated lagged deviation from PPP. I f 7 1 is positive and significantly different
from zero, then the domestic price level tends to move to restore the long-run
PPP relationship. Engle and Granger (1987) and Stock (1987) show that (9) yields
consistent estimates o f a 0 and a ~Engle . ~and Granger (1987) show the two-step
procedure to estimate (8) and (9) is consistent and asymptotically e f f i ~ i e n t . ~
The domestic price level could and does adjust toward its PPP level whether
exchange rates are fixed or flexible. It is expected that the parameters o f the
error-correction model o f inflation would be different under different exchange
rate regimes. The more the authorities allow the nominal exchange rate to change,
the less the required adjustment o f domestic prices to maintain PPP. Thus equation
(8) is estimated separately over time periods that correspond to fixed and flexible
exchange rate regimes.
I f the exchange rate is flexible the cointegrating regression is written as

The error-correction model o f the change in the exchange rate is

The term following Y I is the error-correction term to force the adjustment o f the
nominal exchange rate towards its PPP value. Equation ( 1 1 ) can be estimated only
from periods o f data when nominal exchange rates were allowed to move.
I f PPP holds and a1 = dl = 1, the error-correction terms in ( 8 ) and ( 1 1 ) are dif-
ferent normalizations on the same equilibrium relationship. In practice, the residuals
from regressions ( 9 ) and (10) are nearly equal in magnitude and opposite in sign.
Both error-correction models, (8) and ( 1 I), are estimated with the residuals from
6 Stock (1987) shows this property holds even if both variables are endogenous. Bannerjee et al.
(1986) show the simultaneous-equations bias in the co-integrating regression is small as long as
the R~ in the regression is high.
7 Stock (1987) compares, using Monte Carlo analysis, the one-step non-linear squares estimate
of (8) with the two-step estimates using (8) and (9). Both estimates are consistent and produce
appropriate standard errors. Stock's results and those of Bannejee et al. (1986) suggest the one-
step estimate may he slightly better. The one-step estimator is not available in this study because
the error-correction mechanisms, although not the estimation of deviations from PPP, differ by
exchange rate regime.
 Co-integration, error correction 845

(9).This allows the co-integrating regression to be estimated from the longest pos-
sible time series, as suggested by Hendry (1986), and allows the error-correction
mechanism to vary across different nominal exchange rate regimes. Use o f the
residuals from (9) as error-correction terms in both the error-correction model o f
prices (8) and exchange rates ( 1 1 ) means 71, the coefficient on the error-correction
term, should have opposite signs in ( 8 ) and ( 1 1). I f the residual in (9) is positive,
Canadian goods are expensive relative to their foreign counterparts. I f the residual
in ( 9 ) is positive, Canadian prices tend to fall in (8) or the exchange rate (Canadian
cents per u.s dollar) in ( 1 1 ) tends to depreciate.

1V. RESULTS

This study produces substantial evidence in favour o f PPP between Canada and the
United States. Tests for co-integration reveal a long-run relationship between na-
tional price levels expressed in the same currency. In the associated error-correction
model o f the change in the nominal exchange rate when exchange rates are ad-
justable, the deviation from PPP, the error-correction term, is consistently significant
and of the correct sign. Whether nominal exchange rates are fixed or flexible, the
error-correction term usually plays a role in the determination o f the rate o f change
o f Canadian prices. The error-correction'mechanisms, as anticipated, do vary by
exchange rate regime. The structure o f the data set and its division into periods o f
fixed and flexible exchange rates is considered below.

Structure of data
Two data sets are analysed, an annual data set (1870-1986) and a quarterly data
set (1950.3-1986.4).8 The annual data is divided into three periods: a brief floating
exchange rate period 1870-8 (the greenback era), a longer fixed exchange rate
period 1879-1913 (the gold standard), and a period o f adjustable nominal exchange
rates 1914-86. The 1914-86 time period does include subperiods where nominal
exchange rates were fixed for up to a decade.
The quarterly data 1950.3-1986.4 are also divided into three periods. Canada
began its first post-war flexible exchange rate regime on 30 September 1950. The
error-correction models o f changes in the exchange rate or the price level are es-
timated starting in 1952.1, that is, after the complete removal o f wartime foreign
exchange controls on 14 December 1951.9 Although the Canadian dollar was not
formally fixed until 2 May 1962, the budget speech o f 20 December 1960 in-
cluded a clear announcement that any further appreciation o f the Canadian dollar
would be resisted and that a substantial depreciation o f the Canadian dollar was
the longer-tenn goal o f the government. In the present study the end o f flexible

8 Details concerning sources are found in the data appendix.

9 A description of the purpose and mechanism of foreign exchange control is found in Rasminsky
(1941) or Gibbons (1953). It is clear from the Annual Reports, Foreign Exchange Control Board
(1947 through 1951) that foreign exchange controls and other restrictions on imports were almost
completely relaxed by the beginning of the flexible exchange rate regime on 30 September 1950.
846 David R. Johnson

exchange rates is set at 1960.4.1° The period of fixed exchange rates is then set
from 1961.1 to 1970.1. The recent period of flexible exchange rates is from 1970.2
to 1986.4." Although the error-correction models of inflation or exchange rate
change are estimated using the subperiods corresponding to the different exchange
rate regimes, tests for co-integration are carried out for the longest possible time
periods spanning the different exchange rate regimes.

Tests for co-integration

Table 1 presents the estimates of the co-integrating regressions (9) and (10) for the
annual data, two subperiods of the annual data and the quarterly data.
Engle and Granger (1987) outline and provide critical values, included in table
1, for three test statistics that serve as a 'rough guide' in testing the null hypothesis
that a bivariate vector times series is co-integrated. The intuition for these statistics
is as follows. If PPP holds the estimated deviations from PPP, the residual e, from
the co-integrating regression (9) or (lo), cannot be non-stationary.12 If the residual
were non-stationary the real exchange rate would drift away from its PPP value.
A Durbin-Watson (DW)statistic that is close to zero indicates the hypothesis that
deviations from PPP are non-stationary cannot be rejected. A second test statistic is
based on the t-statistic (DF)of the coefficient p in the regression

where e, is the residual from the co-integrating regression (9) or (10). The estimate
of p must be large enough to reject the hypothesis that the residuals from the PPP
regression are non-stationary. A similar t-statistic (the ADF statistic) is constructed
from the estimate of p in

Ae, = per- 1 +

The use of the ADF statistic is recommended over the DF statistic if the coefficients si
are non-zero or there is a possibility of seasonal dynamics. In table 1 the appropriate
statistic is presented for each co-integrating regression.
There is some evidence that the annual times series are co-integrated. Equation
(9) is estimated using the full data period, 1870-1986, the gold standard period,

10 Plumptre (1970, 3) states: 'The period of float came to an end de facto in 1961 when the Cana-
dian government took a position in regard to the (foreign exchange) rate and de jure in May
1962.' Wonnacott (1965), Shepherd (1973) and, of course, Plumptre (1977) amve at the same
conclusion.
11 The substantive results for the error-correction models in tables 3 and 4 are not changed if
1950.3-1962.1 is used as the early period of flexible rates and 1962.2-1970.1 is used as the
period of fixed exchange rates. I am grateful to an anonymous referee for pointing out the neces-
sity of being careful in dating the transition from flexible to fixed exchange rates between 1960
and 1962.
12 Pippenger (1986) employs a similar methodology in analysing the residuals from PPP using spec-
tral analysis.
 Co-integration, error correction 847

TABLE 1
Co-integrating regressions

Equation Period Estimatea Statisticsb

1.I 1870-86 In P, = -4.21 + 0.962 In (E,P*,) R2 = 0.994

DW = 0.432
DF = -3.79

1.2 1879-1913 In P, = -3.83 + 0.913 In (E,P*,) R2 = 0.802

DW = 0.693
DF = -2.60

1.3 1914-86 In P, = -4.42 + 0.985 In (E,P*,) R2 = 0.992

OW = 0.430
DF = -2.99

1.4 1914-86 In E, = 4.58 + 0.674 In (P,/P*,) RZ = 0.430

DW = 0.366
A D F ( ~=) -2.36
1.5 1950.3-86.4 In P, = -4.35 + 0.974 In (E,P*,) R2 = 0.988
OW = 0.263
ADF(4) = -4.15
1.6 1950.3-86.4 In E, = -4.59 + 0.942 In (P,/P*,) R2 = 0.635
OW = 0.232
A D F ( ~=) -3.77

a No standard errors are reported in the co-integrating regressions. Computed standard errors in regres-
sions involving non-stationary variables are understated. Regressions (1.2), (1.3), and (1.5) are used in
generation of the deviations from PPP, the error-correction terms.
b Critical values (1, 5, and 10 per cent) to reject the null hypothesis of no co-integration in a bivariate
time series are provided in Granger and Engle (1987) and reproduced below, Rejection of the null
hypothesis indicates support of the PPP proposition.
Statistic Model A Model B
1% 5% 10% 1% 5% 10%
DW 0.511 0.386 0.322 0.455 0.282 0.204
DF 4.07 3.3'7 3.03 3.90 3.05 2.71
ADF(S) 3.77 3.17 2.84 3.73 3.17 2.91
Model El allows for more extensive dynamics, including seasonal effects. The three test statistics are
discussed briefly in the text and extensively in Engle and Granger (1987). S is the number of lags in
regression ( ) for the ADF statistic.

1879-1913, and the period of adjustable exchange rates, 1914-1986. In each case
the point estimate of a is near unity, as predicted by PPP. The ow statistic rejects the
null hypothesis of no co-integration in all three cases at 5 per cent or higher. The
DF statistic rejects the null hypothesis of no co-integration at 5 per cent using the
full sample and at nearly 10 per cent using the 1914-86 subsample. Equation (10)
is estimated (see equation (1.4)) using the 1914-86 period of adjustable exchange
rates. Only the DW statistic provides any suggestion of co-integration.
Evidence for co-integration is substantially stronger in the post-1950 quarterly
data. The estimates of both a and cc' from equations (9) and (10) (see equations
(1.5) and (1 6 ) ) are close to unity, as predicted by PPP. Both ADF statistics reject the
848 David R. Johnson

i
20

15
-------
.........
Residuals from P.quauon I l
u , O u 0 . 3 n

Deviations
from PPP
(percent)

FIGURE 1 Deviations from purchasing power parity, 1870-1986 (SOURCE: Residuals from co
integrating regressions in table 1

null hypothesis of no co-integration at 1 per cent, both Durbin-Watson statistics

reject the null at 10 per cent.13
Figure 1 plots the residuals from equations (1.1), (1.3), and (1.5). All three co-
integrating regressions generate a similar a pattern of residuals. The lagged values
of these residuals are used as the error-correction terms.

Annual data
Table 2 presents estimates of the error-correction models using annual data. During
the gold standard the nominal exchange rate is fixed. Only prices are able to adjust
to maintain PPP. Equation (2.1) models the adjustment in domestic inflation over
the 1880-1913 period. The error-correction term is of the correct sign, although
it is not statistically significant. This insignificant error-correction coefficient is
consistent with the weak rejection of the null hypothesis of no co-integration in
equation (1.2). A strong mechanism for the maintenance of PPP does not emerge in
the gold standard period.
It is necessary to ensure the residuals in equation (2.1) and all other error-

13 Tests following Fuller (1976), Dickey and Fuller (1981), or Hasza and Fuller (1982), as appro-
priate, were conducted on the logarithm of the real exchange rate. The null hypothesis that the
logarithm of the real exchange rate contains a unit root cannot be rejected for the annual samples
from 1879-1986 and 1879-1913. The null hypothesis is rejected (i,, = -2.97) for the 1914-86
subperiod. The null hypothesis is also rejected for the full sample of quarterly data (1950.3-
1986.4) (i = -4.023) and is rejected only at a 0.10 level of significance (^7, = -2.70) for
the quarterly subsample from 1973-86. It is noteworthy that most studies that reject PPP through
the inability to reject the null hypothesis of a unit root in the logarithm of the real exchange rate
restrict themselves to the post-1973 period. Critical values of .i, are found in Fuller (1976). The
results of tests for the unit root correspond exactly to the results of tests for co-integration in
table 1.
 Co-integration, error correction 849

TABLE 2
Error-correction models, annual data

Equation Period E~timate",~ CI regressionc

a Standard errors in parentheses

b The Q + ( p ) statistic is distributed F ( p , T - p - k) on the null hypothesis of p zero autocor-
relations in the residuals from a regression including lagged dependent variables. T is the number of
observations and k the number of regressors.
c The error-correction term EC, is the lagged residual from the cointegrating ( a ) regression in table 1.
A significant coefficient on this term indicates that the deviation from PPP affects the path of the
nominal exchange rate or the Canadian inflation rate.

correction models are serially uncorrelated. A Lagrange multiplier test statistic

Q + ( p ) (Harvey 1981) is used to test the null hypothesis that the first p autocorre-
lations are zero. The Q + ( p ) statistic is recommended when the regression involves,
as all error-correction models do, a lagged dependent variable. This test statistic is
reported in each case.
Over the 1914-86 period there were a number of changes in the nominal ex-
change rate. An error-correction model of exchange rate change, equation (2.2),
contains an error-correction term which is correctly signed and statistically sig-
nificant. This is evidence to add to the somewhat mixed evidence in favour of
co-integration from equations (1.1), (1.3), and (1.4). Equation (2.2) suggests that
a 10 per cent deviation from PPP is followed, on average, by a 2 per cent annual
change in the exchange rate. The model can also be interpreted to suggest that the
Canadian monetary authorities took deviations from PPP into account in intervention
to affect or fix a nominal exchange rate.
Equation (2.3) is the error-correction representation of the rate of change in
the Canadian price level over the 1914-86 period. The error-correction term is
incorrectly signed, small in absolute magnitude, and statistically insignificant.
The combination of (2.2) and (2.3) leads to an interesting and useful interpreta-
tion concerning the mechanisms through which PPP is maintained. Lags of foreign
inflation and domestic inflation are not significant when added to equation (2.2),
the error-correction model of the change in the exchange rate. Such variables could,
but need not, enter into this model (see equation (7)). The importance of domes-
850 David R. Johnson

tic or foreign inflation for exchange rate change is completely captured by the
error-correction term. Similarly lagged exchange rate change and lagged foreign
inflation do not enter equation (2.3), the error-correction model of domestic infla-
tion. If nominal exchange rates are truly adjustable, the domestic price process is
independent of the foreign price p r o ~ e s s . ' ~

Quarterly data
The estimates of quarterly error-correction models appear in tables 3 and 4. They
provide further insight into the fact that PPP can be maintained in different ways
over different exchange rate regimes. Equations (3.1) and (3.2) model nominal
exchange rate change in the early (1950s) and later (after 1970) flexible exchange
rate regimes. The coefficient on the error-correction term is significant in both
periods but twice as large in the second period of floating rates. A stronger response
of the exchange rate to deviations from PPP is an important feature of the second
period of flexible exchange rates.
Further support for the importance of changes in regime is found when error-
correction models of inflation are estimated separately over the two flexible ex-
change rate periods. Two error-correction models of inflation are estimated for
each period of flexible exchange rates. Equations (3.5) and (3.6) allow for the
possibility of seasonality in inflation. Equations (3.3) and (3.4) do not explicitly
allow for seasonality, although only the fourth lag of inflation is significant in
either equation. The equations reveal that the mechanism for the maintenance of
PPP through variations in the domestic inflation rate did vary over the two flexible
exchange rate regimes. In the 1950s float, where the adjustment in the nominal
exchange rate is weaker, the role of error-correction in the determination of do-
mestic inflation is stronger. The coefficient on the error-correction term in equation
(3.3) is twice as large as in equation (3.4). In the latter equation, the coefficient on
the error-correction term is not statistically significant. The coefficient pattern is
similar in comparing equations (3.5) and (3.6). In equation (3.6), the coefficient on
the error-correction term is statistically significant but only one-half the magnitude
of the equivalent coefficient in equation (3.5). In the post-1970 flexible exchange
rate period, the nominal exchange rate was allowed more freedom to move and the
role of error-correction in the determination of domestic inflation is much smaller.
This corresponds closely to the textbook model of a pure flexible exchange rate
regime.
If the nominal exchange rate is fixed, only the domestic inflation rate is able
to adjust to maintain the PPP relationship. Error-correction models of inflation es-
timated using data from the fixed exchange rate period are found in table 4. In-
tervention to reduce and then fix the value of the Canadian dollar between 1960
and 1962 involved, according to figure 1, a significant depreciation of both the real

14 Engle and Granger (1987) note the absence of a significant error-correction term in one of the
error-correction equations corresponds to weak exogeneity of that variable. Thus if exchange rates
are truly flexible, domestic inflation is weakly exogenous. The long annual data period includes a
substantial depreciation of the Canadian dollar.
 Go-integration, error correction 85 1

TABLE 3
Error-correction models, quarterly data, flexible exchange ratesa

Equation Period Estimateb '

a The flexible exchange rate period in the 1950s is dated from the removal of foreign exchange controls
on 14 December 1951 to the beginning of active intervention in foreign exchange markets as announced
on 20 December 1960. See text for further discussion.
b Standard errors in parentheses
c The Q + (p) statistic is distributed F ( p , T - p - k) on the null hypothesis of p zero autocorrelations
in the residuals from a regression including lagged dependent variable. 7' is the number of observa-
tions and k the number of regressors.
d The error-correction term EC, is the lagged residual (fourth lag of the residual in equations (3.5) and
(3.6)) from the co-integrating regression in table 1. A significant coefficient on this term indicates that
the deviation from PPP affects the path of the nominal exchange rate or the Canadian inflation rate.

and nominal exchange rate. The error-correction term that responds to that under-
valuation is large and strongly significant in both equations in table 4. Canadian
inflation during the 1960s was accelerated by the return towards the PPP relation-
ship. It is also noteworthy that in equation (4.2), the seasonally adjusted model of
Canadian inflation, lagged American inflation has a statistically significant impact,
independent of the error-correction mechanism, on Canadian inflation. No such ef-
fects could be found in the error-correction models of domestic price change or
exchange rate change when exchange rates were flexible. The direct influence of
foreign inflation on domestic inflation is part of the textbook model of a fixed
exchange rate regime.
852 David R. Johnson

TABLE 4
Error-correction models, quarterly data, fixed exchange ratesa

Equation Period Estimateb,' CI regressiond

4.1 1961.1-70.1 A In PI = 0.576 A In PI_, - 0.065 EC, 1.5

(0.131) (0.018)
R2 = 0.354 Q + ( 4 ) = 2.14
4.2 1961.1-70.1 A4 In P, = 0.376A4 In P,-, + 0.324A4 In P,-, 1.5
(0.170) (0.159)
+ 0.182A41n P*,.., - 0 . 0 8 8 ~ ~ ~
(0.084) (0.031)
R2 = 0.873 Q + ( 4 ) = 1.36

a The period of fixed exchange rates is dated from 20 December 1960, when the authorities announced
active intervention to reduce the value of the Canadian dollar. Exchange rates were officially fixed on
2 May 1962. See text for further discussion.
b Standard errors in parentheses
c The Q + ( p ) statistic is distributed F ( p , T - p - k ) on the null hypothesis of p zero autocorrelations
in the residuals from a regression including lagged dependent variable. T is the number of observations
and k the number of regressors.
d The error-correction term EC, is the lagged residual (fourth lag of the residual in equation 4.2) from the
co-integrating regression in table 1. A significant coefficient on this term indicates that the deviation
from PPP affects the path of the Canadian inflation rate.

IV. CONCLUSIONS

The more extensive dynamic framework adopted in this study uncovers significant
new evidence in favour of PPP as the long-run relationship between Canadian prices,
American prices, and the Canadian dollar / American dollar exchange rate. The
relevant time series are co-integrated after 1950, and there is some evidence in
favour of co-integration over the post-1914 period. Error-correction terms play a
significant role in modelling both the change in the nominal exchange rate and, to
a slightly smaller extent, the change in the price level.
The results also yield useful insights into the mechanism for maintenance of
PPP. This mechanism clearly depends on the choice of exchange rate regime. Over
the entire 1914-86 period, and to some extent after 1970, the results of this study
suggest that a fully adjustable nominal exchange rate allows Canadian inflation
to be independent of the American inflation. After 1950 three nominal exchange
rate regimes were implemented. During the period of fixed nominal exchange rates
during the 1960s PPP is maintained through adjustment of Canadian prices toward
their PPP level. When the exchange rate was fixed, Canadian inflation responded
directly to American inflation. The two post-1950 periods of flexible exchange
rates were different. The period of exchange rate flexibility after 1970 corresponded
more closely to the textbook model of flexible exchange rates. The flexible nominal
exchange rate responded strongly to deviations from PPP after 1970. There was only
a small response of Canadian inflation to deviations from PPP in the 1970s rather
than the larger response in the 1950s. In the 1950s the nominal exchange rate,
 Go-integration, error correction 853

though flexible, moved less in response to deviations from PPP than in the 1970s.
The implementation of exchange rate policy and the choice of exchange rate regime
have a strong effect on the mechanism through which PPP is maintained and thus a
strong effect on the domestic inflation process.

APPENDIX: DATA SOURCES

Annual data
GNP Deflator - United States
1870-1983 Nathan S. Balke and Robert J. Gordon (1986) 'Historical data.' In
Robert J. Gordon, ed., The American Business Cycle: Continuity and Change
(Chicago: University of Chicago Press)
1983-86 Bureau of Economic Analysis, us. Department of Commerce (February
1987) Survey of Current Business

GNP Deflator - Canada

1870-1926 M.C. Urquhart (1986) 'New estimates of gross national product,
Canada, 1870-1926: some implications for Canadian developments.' In Stanley
L. Engerman and Robert E. Gallman, eds, Long-Term Factors in American Eco-
nomic Growth (Chicago: University of Chicago Press)
1926-86 CANSIM Variable D40672.

Exchange Rate - Canadian cents per u.s. dollar

187C-78 James K. Kindahl (1961) 'Economic factors in specie resumption: the
United States 1865-79.' Journal of Political Economy 69, 30-48. This source
provides the greenback price of a gold us. dollar and hence the exchange rate
per Canadian dollar.
1879-1913 The us. and Canadian dollar are defined as and fully convertible into
the same amount of gold. The exchange rate is unity.
1914-45 Column J562 (1983) 'Average daily noon exchange rate.' In F.H. Leacy,
ed. Historical Statistics of Canada, Second Edition (Ottawa: Statistics Canada)
1946-86 The annual average of CANSIM Variable B40001, a monthly exchange rate
series constructed from an average of daily observations.

Quarterly data
GNP Deflator - United States
1947.1-1950.4 Federal Reserve Bank of St Louis
1951.1-1984.4 Bureau of Economic Analysis, us. Department of Commerce
(February 1986) A Supplement to the Survey of Current Business
1985.1-1986.4 Bureau of Economic Analysis, u.s. Department of Commerce
(February 1987) Survey of Current Business

GNP Deflator - Canada

1947.4-1986.4 CANSIM Variable D40012 (current dollar GNP)divided by CANSIM
Variable D40561 (constant dollar GNP)
854 David R. Johnson

Exchange Rate - Canadian cents per u.s. dollar

1946.1-1986.4 Quarterly Average of CANSIM Variable B40001, a monthly average
of daily observations

REFERENCES

Adler, M. and B. Lehman (1983) 'Deviations from purchasing power parity in the long
run.' Journal of Finance 38, 1471-87
Baillie, R.T. and D.D. Selover (1987) 'Cointegration and models of exchange rate deter-
mination.' International Journal of Forecasting 3, 43-51
Bannerjee, A., J.J. Dolado, D.F. Hendry, and G.W. Smith (1986) 'Exploring equilibrium
relationships in econometrics through static models: some Monte Carlo evidence.'
Oxford Bulletin of Economics and Statistics 48, 253-77
Campbell, John Y. (1987) 'Does saving anticipate declining labour income? An alternative
test of the permanent income hypothesis, Econometrica 55, 1249-74.
Campbell, John Y. and Robert J. Shiller (1987) 'Cointegration and tests of present value
models.' Journal of Political Economy 95, 1062-88
Corbae, Dean and Sam Ouliaris (1988) 'Cointegration and tests of purchasing power
parity.' Review of Economics and Statistics.' L X X , 508-11
Cumby, R.E. and M. Obstfelt (1984) 'Inernational interest rate and price level linkages:
a review of recent evidence.' In J.F.O. Bilson and R.C. Marston, eds, Exchange Rate
Theory and Practice (Chicago: University of Chicago Press)
Davidson, James E.H., David F. Hendry, Frank Srba and Stephen Yeo (1978) 'Economet-
ric modeling of the aggregate time-series relationship between consumer's expenditure
and income in the United Kingdom. Economic Journal 88, 661-92
Dickey, David A. and Wayne A. Fuller (1981) 'Likelihood ratio statistics for autoregres-
sive time series with a unit root.' Econometrica 49, 1057-72
Edison, Hali J. (1987) 'Purchasing power parity in the long run: a test of the dollar/pound
exchange rate (1890-1978)' Journal of Money, Credit and Banking 19, 376-87
Enders, Walter (1988) 'Arima and cointegration tests of PPP under fixed and flexible
exchange rate regimes.' Review of Economics and Statistics 70, 504-8
Engle, R.F. and C.W.G. Granger (1987) 'Cointegration and error-correction: representa-
tion, estimation and testing.' Econometrica 55, 251-76
Foreign Exchange Control Board, 1946 through 1951, Annual Report to the Minister of
Finance (Ottawa: King's (Queen's) Printer)
Frankel, J.A. (1986) 'International capital mobility and crowding-out in the u s economy:
imperfect integration of financial markets or goods markets?' In R.W. Hafer, ed., How
Open is the U.S. Economy? (Lexington, M A : Lexington Books)
- (1981) 'Flexible exchange rates, prices and the role of 'news': lessons from the
1970's.' Journal of Political Economy 89, 665-705
Fuller, Wayne A. (1976) Introduction to Statistical Times Series (New York: Wiley)
Gibbons, A.O. (1953) 'Foreign exchange control in Canada, 1939-51.' Canadian Journal
of Political Science and Economics 19, 35-54
Granger, C.W.G. and P. Newbold (1986) Forecasting Economic Time Series, Second
Edition (New York: Academic Press)
Hall, S.G. (1986) 'An application of the Granger and Engle two-step estimation procedure
to United Kingdom aggregate wage data.' Oxford Bulletin of Economics and Statistics
48, 229-39
Harvey A.C. (1981) The Econometric Analysis of Time Series (New York: Wiley)
Hasza, David P. and Wayne A. Fuller (1982) 'Testing for nonstationary parameter specifi-
cations in seasonal time series models.' The Annals of Statistics 10, 1209-16
 Go-integration, error correction 855

Hendry, D.F. (1986) 'Econometric modelling with cointegrated variables: an overview.'

Oxford Bulletin of Economics and Statistics 48, 201-12
Huizinga, John (1987) 'An empirical investigation of the long-run behaviour of real
exchange rates.' In K. Brunner and A.H. Meltzer, eds, Empirical Studies of Velocity,
Real Exchange Rates, Unemployment and Productivity (Amsterdam: North-Holland)
Jenkinson, T.G. (1986) 'Testing neo-classical theories of labour demand: an application of
cointegration techniques.' Oxford Bulletin of Economics and Statistics 48, 241-51
Katseli-Papaefstratiou, L.T. (1979) 'The re-emergence of purchasing power parity doctrine
in the 1970's.' Special Papers in International Economics 13, Princeton University
Levich, Richard M. (1985) 'Empirical studies of exchange rates: price behaviour, rate
determination and market efficiency.' In Ronald W. Jones and Peter B. Kenen, eds,
Handbook of International Economics vol. 11 (Amsterdam: North-Holland)
Mark, Nelson C. (1987) 'Real and nominal exchange rates in the long run: an empirical
investigation.' Mimeo, Ohio State University
Meese, Richard A. and Kenneth J. Singleton (1982) 'On unit roots and the empirical
modelling of exchange rates.' Journal of Finance 37, 1029-35
Mishkin, F.S. (1984) 'Are real interest rates equal across countries: an empirical investiga-
tion of international parity conditions.' Journal of Finance 39, 1345-58
Nelson, Charles R. and Charles I. Plosser (1982) 'Trends and random walks in'macroeco-
nomic time series.' Journal of Monetary Economics 10, 1 3 9 4 2
Officer, Lawrence (1976) 'The purchasing power parity theory of exchange rates: a review
article.' IMF Staff Papers 23, 1 4 0
Pippenger, John (1982) 'PPP:an analysis of predictive error.' This JOURNAL 15, 335-46
Plumptre, A.F.W. (1970) 'Exchange rate policy: experience with Canada's floating rate.'
Essays in International Finance No. 81, Princeton University
- (1977) Three Decades of Decision: Canada and the World Moneta~ySystem 1944-75
(Toronto: McClelland and Stewart)
Rasminsky, Louis (1941) 'Foreign exchange control: purpose and methods.' In J.F. Parkin-
son, ed., Canadian War Economics (Toronto: University of Toronto Press)
Roll, R. (1979) 'Violations of purchasing power parity and their implications for efficient
international commodity markets.' In M. Sarnat and G.P. Szego, eds, International
Finance and Trade (Cambridge, M A : Ballinger)
Shepherd, Sidney A. (1973) Foreign Exchange and Foreign Trade in Canada: An Outline,
Fourth Edition (Toronto: University of Toronto Press)
Stock, James H. (1987) 'Asymptotic properties of least squares estimators of cointegrating
vectors.' Econometrica 55, 1035-56
Stock, James H. and Mark Watson, 'Testing for common trends.' Journal of the American
Statistical Association, forthcoming.
Taylor, Mark P. (1988) 'An empirical examination of long-run purchasing power parity
using cointegration techniques.' Applied Economics 20, 1369-81
Wasserfallen, Walter (1986) 'Non-stationarities in macroeconomic time series - further
evidence and implications.' This JOURNAL 19, 498-510
Wonnacott, Paul (1965) The Canadian Dollar 1948-1962 (Toronto: University of Toronto
Press)
http://www.jstor.org

LINKED CITATIONS
- Page 1 of 6 -

You have printed the following article:

Co-Integration, Error and Purchasing Power Parity between Canada and the United States
David R. Johnson
The Canadian Journal of Economics / Revue canadienne d'Economique, Vol. 23, No. 4. (Nov.,
1990), pp. 839-855.
Stable URL:
http://links.jstor.org/sici?sici=0008-4085%28199011%2923%3A4%3C839%3ACEAPPP%3E2.0.CO%3B2-F

This article references the following linked citations. If you are trying to access articles from an
off-campus location, you may be required to first logon via your library web site to access JSTOR. Please
visit your library's website or contact a librarian to learn about options for remote access to JSTOR.

[Footnotes]

2
Co-Integration and Error Correction: Representation, Estimation, and Testing
Robert F. Engle; C. W. J. Granger
Econometrica, Vol. 55, No. 2. (Mar., 1987), pp. 251-276.
Stable URL:
http://links.jstor.org/sici?sici=0012-9682%28198703%2955%3A2%3C251%3ACAECRE%3E2.0.CO%3B2-T

2
Does Saving Anticipate Declining Labor Income? An Alternative Test of the Permanent
Income Hypothesis
John Y. Campbell
Econometrica, Vol. 55, No. 6. (Nov., 1987), pp. 1249-1273.
Stable URL:
http://links.jstor.org/sici?sici=0012-9682%28198711%2955%3A6%3C1249%3ADSADLI%3E2.0.CO%3B2-J

2
Cointegration and Tests of Present Value Models
John Y. Campbell; Robert J. Shiller
The Journal of Political Economy, Vol. 95, No. 5. (Oct., 1987), pp. 1062-1088.
Stable URL:
http://links.jstor.org/sici?sici=0022-3808%28198710%2995%3A5%3C1062%3ACATOPV%3E2.0.CO%3B2-6

NOTE: The reference numbering from the original has been maintained in this citation list.
http://www.jstor.org

LINKED CITATIONS
- Page 2 of 6 -

3
Non-Stationarities in Macro-Economic Time Series -- Further Evidence and Implications
Walter Wasserfallen
The Canadian Journal of Economics / Revue canadienne d'Economique, Vol. 19, No. 3. (Aug.,
1986), pp. 498-510.
Stable URL:
http://links.jstor.org/sici?sici=0008-4085%28198608%2919%3A3%3C498%3ANIMTS-%3E2.0.CO%3B2-S

3
On Unit Roots and the Empirical Modeling of Exchange Rates
Richard A. Meese; Kenneth J. Singleton
The Journal of Finance, Vol. 37, No. 4. (Sep., 1982), pp. 1029-1035.
Stable URL:
http://links.jstor.org/sici?sici=0022-1082%28198209%2937%3A4%3C1029%3AOURATE%3E2.0.CO%3B2-V

4
Likelihood Ratio Statistics for Autoregressive Time Series with a Unit Root
David A. Dickey; Wayne A. Fuller
Econometrica, Vol. 49, No. 4. (Jul., 1981), pp. 1057-1072.
Stable URL:
http://links.jstor.org/sici?sici=0012-9682%28198107%2949%3A4%3C1057%3ALRSFAT%3E2.0.CO%3B2-4

5
Deviations from Purchasing Power Parity in the Long Run
Michael Adler; Bruce Lehmann
The Journal of Finance, Vol. 38, No. 5. (Dec., 1983), pp. 1471-1487.
Stable URL:
http://links.jstor.org/sici?sici=0022-1082%28198312%2938%3A5%3C1471%3ADFPPPI%3E2.0.CO%3B2-M

5
Are Real Interest Rates Equal Across Countries? An Empirical Investigation of International
Parity Conditions
Frederic S. Mishkin
The Journal of Finance, Vol. 39, No. 5. (Dec., 1984), pp. 1345-1357.
Stable URL:
http://links.jstor.org/sici?sici=0022-1082%28198412%2939%3A5%3C1345%3AARIREA%3E2.0.CO%3B2-R

6
Asymptotic Properties of Least Squares Estimators of Cointegrating Vectors
James H. Stock
Econometrica, Vol. 55, No. 5. (Sep., 1987), pp. 1035-1056.
Stable URL:
http://links.jstor.org/sici?sici=0012-9682%28198709%2955%3A5%3C1035%3AAPOLSE%3E2.0.CO%3B2-5

NOTE: The reference numbering from the original has been maintained in this citation list.
http://www.jstor.org

LINKED CITATIONS
- Page 3 of 6 -

7
Asymptotic Properties of Least Squares Estimators of Cointegrating Vectors
James H. Stock
Econometrica, Vol. 55, No. 5. (Sep., 1987), pp. 1035-1056.
Stable URL:
http://links.jstor.org/sici?sici=0012-9682%28198709%2955%3A5%3C1035%3AAPOLSE%3E2.0.CO%3B2-5

13
Likelihood Ratio Statistics for Autoregressive Time Series with a Unit Root
David A. Dickey; Wayne A. Fuller
Econometrica, Vol. 49, No. 4. (Jul., 1981), pp. 1057-1072.
Stable URL:
http://links.jstor.org/sici?sici=0012-9682%28198107%2949%3A4%3C1057%3ALRSFAT%3E2.0.CO%3B2-4

13
Testing for Nonstationary Parameter Specifications in Seasonal Time Series Models
David P. Hasza; Wayne A. Fuller
The Annals of Statistics, Vol. 10, No. 4. (Dec., 1982), pp. 1209-1216.
Stable URL:
http://links.jstor.org/sici?sici=0090-5364%28198212%2910%3A4%3C1209%3ATFNPSI%3E2.0.CO%3B2-E

14
Co-Integration and Error Correction: Representation, Estimation, and Testing
Robert F. Engle; C. W. J. Granger
Econometrica, Vol. 55, No. 2. (Mar., 1987), pp. 251-276.
Stable URL:
http://links.jstor.org/sici?sici=0012-9682%28198703%2955%3A2%3C251%3ACAECRE%3E2.0.CO%3B2-T

References

Deviations from Purchasing Power Parity in the Long Run

Michael Adler; Bruce Lehmann
The Journal of Finance, Vol. 38, No. 5. (Dec., 1983), pp. 1471-1487.
Stable URL:
http://links.jstor.org/sici?sici=0022-1082%28198312%2938%3A5%3C1471%3ADFPPPI%3E2.0.CO%3B2-M

NOTE: The reference numbering from the original has been maintained in this citation list.
http://www.jstor.org

LINKED CITATIONS
- Page 4 of 6 -

Does Saving Anticipate Declining Labor Income? An Alternative Test of the Permanent
Income Hypothesis
John Y. Campbell
Econometrica, Vol. 55, No. 6. (Nov., 1987), pp. 1249-1273.
Stable URL:
http://links.jstor.org/sici?sici=0012-9682%28198711%2955%3A6%3C1249%3ADSADLI%3E2.0.CO%3B2-J

Cointegration and Tests of Present Value Models

John Y. Campbell; Robert J. Shiller
The Journal of Political Economy, Vol. 95, No. 5. (Oct., 1987), pp. 1062-1088.
Stable URL:
http://links.jstor.org/sici?sici=0022-3808%28198710%2995%3A5%3C1062%3ACATOPV%3E2.0.CO%3B2-6

Cointegration and Tests of Purchasing Power Parity

Dean Corbae; Sam Ouliaris
The Review of Economics and Statistics, Vol. 70, No. 3. (Aug., 1988), pp. 508-511.
Stable URL:
http://links.jstor.org/sici?sici=0034-6535%28198808%2970%3A3%3C508%3ACATOPP%3E2.0.CO%3B2-K

Econometric Modelling of the Aggregate Time-Series Relationship Between Consumers'

Expenditure and Income in the United Kingdom
James E. H. Davidson; David F. Hendry; Frank Srba; Stephen Yeo
The Economic Journal, Vol. 88, No. 352. (Dec., 1978), pp. 661-692.
Stable URL:
http://links.jstor.org/sici?sici=0013-0133%28197812%2988%3A352%3C661%3AEMOTAT%3E2.0.CO%3B2-J

Likelihood Ratio Statistics for Autoregressive Time Series with a Unit Root
David A. Dickey; Wayne A. Fuller
Econometrica, Vol. 49, No. 4. (Jul., 1981), pp. 1057-1072.
Stable URL:
http://links.jstor.org/sici?sici=0012-9682%28198107%2949%3A4%3C1057%3ALRSFAT%3E2.0.CO%3B2-4

Purchasing Power Parity in the Long Run: A Test of the Dollar/Pound Exchange Rate
(1890-1978)
Hali J. Edison
Journal of Money, Credit and Banking, Vol. 19, No. 3. (Aug., 1987), pp. 376-387.
Stable URL:
http://links.jstor.org/sici?sici=0022-2879%28198708%2919%3A3%3C376%3APPPITL%3E2.0.CO%3B2-O
NOTE: The reference numbering from the original has been maintained in this citation list.
http://www.jstor.org

LINKED CITATIONS
- Page 5 of 6 -

Arima and Cointegration Tests of PPP under Fixed and Flexible Exchange Rate Regimes
Walter Enders
The Review of Economics and Statistics, Vol. 70, No. 3. (Aug., 1988), pp. 504-508.
Stable URL:
http://links.jstor.org/sici?sici=0034-6535%28198808%2970%3A3%3C504%3AAACTOP%3E2.0.CO%3B2-T

Co-Integration and Error Correction: Representation, Estimation, and Testing

Robert F. Engle; C. W. J. Granger
Econometrica, Vol. 55, No. 2. (Mar., 1987), pp. 251-276.
Stable URL:
http://links.jstor.org/sici?sici=0012-9682%28198703%2955%3A2%3C251%3ACAECRE%3E2.0.CO%3B2-T

Testing for Nonstationary Parameter Specifications in Seasonal Time Series Models

David P. Hasza; Wayne A. Fuller
The Annals of Statistics, Vol. 10, No. 4. (Dec., 1982), pp. 1209-1216.
Stable URL:
http://links.jstor.org/sici?sici=0090-5364%28198212%2910%3A4%3C1209%3ATFNPSI%3E2.0.CO%3B2-E

On Unit Roots and the Empirical Modeling of Exchange Rates

Richard A. Meese; Kenneth J. Singleton
The Journal of Finance, Vol. 37, No. 4. (Sep., 1982), pp. 1029-1035.
Stable URL:
http://links.jstor.org/sici?sici=0022-1082%28198209%2937%3A4%3C1029%3AOURATE%3E2.0.CO%3B2-V

Are Real Interest Rates Equal Across Countries? An Empirical Investigation of International
Parity Conditions
Frederic S. Mishkin
The Journal of Finance, Vol. 39, No. 5. (Dec., 1984), pp. 1345-1357.
Stable URL:
http://links.jstor.org/sici?sici=0022-1082%28198412%2939%3A5%3C1345%3AARIREA%3E2.0.CO%3B2-R

Purchasing Power Parity: An Analysis of Predictive Error

John Pippenger
The Canadian Journal of Economics / Revue canadienne d'Economique, Vol. 15, No. 2. (May,
1982), pp. 335-346.
Stable URL:
http://links.jstor.org/sici?sici=0008-4085%28198205%2915%3A2%3C335%3APPPAAO%3E2.0.CO%3B2-D

NOTE: The reference numbering from the original has been maintained in this citation list.
http://www.jstor.org

LINKED CITATIONS
- Page 6 of 6 -

Asymptotic Properties of Least Squares Estimators of Cointegrating Vectors

James H. Stock
Econometrica, Vol. 55, No. 5. (Sep., 1987), pp. 1035-1056.
Stable URL:
http://links.jstor.org/sici?sici=0012-9682%28198709%2955%3A5%3C1035%3AAPOLSE%3E2.0.CO%3B2-5

Non-Stationarities in Macro-Economic Time Series -- Further Evidence and Implications

Walter Wasserfallen
The Canadian Journal of Economics / Revue canadienne d'Economique, Vol. 19, No. 3. (Aug.,
1986), pp. 498-510.
Stable URL:
http://links.jstor.org/sici?sici=0008-4085%28198608%2919%3A3%3C498%3ANIMTS-%3E2.0.CO%3B2-S

NOTE: The reference numbering from the original has been maintained in this citation list.