Description
The purpose of this paper is to estimate the long-run relationships and threshold effects
between inflation and economic growth in Mexico.
Journal of Financial Economic Policy
Inflation and Mexican economic growth: long-run relation and threshold effects
W. Adrián Risso Edgar J . Sánchez Carrera
Article information:
To cite this document:
W. Adrián Risso Edgar J . Sánchez Carrera, (2009),"Inflation and Mexican economic growth: long-run
relation and threshold effects", J ournal of Financial Economic Policy, Vol. 1 Iss 3 pp. 246 - 263
Permanent link to this document:
http://dx.doi.org/10.1108/17576380911041728
Downloaded on: 24 January 2016, At: 21:36 (PT)
References: this document contains references to 54 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 1334 times since 2009*
Users who downloaded this article also downloaded:
Ahmad Zubaidi Baharumshah, Siew-Voon Soon, (2014),"Inflation, inflation uncertainty and output growth:
what does the data say for Malaysia?", J ournal of Economic Studies, Vol. 41 Iss 3 pp. 370-386 http://
dx.doi.org/10.1108/J ES-05-2012-0073
Girijasankar Mallik, Anis Chowdhury, (2011),"Effect of inflation uncertainty, output uncertainty and oil
price on inflation and growth in Australia", J ournal of Economic Studies, Vol. 38 Iss 4 pp. 414-429 http://
dx.doi.org/10.1108/01443581111160879
Ming-Yu Cheng, Hui-Boon Tan, (2002),"Inflation in Malaysia", International J ournal of Social Economics,
Vol. 29 Iss 5 pp. 411-425 http://dx.doi.org/10.1108/03068290210423532
Access to this document was granted through an Emerald subscription provided by emerald-srm:115632 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for
Authors service information about how to choose which publication to write for and submission guidelines
are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as
providing an extensive range of online products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee
on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.
*Related content and download information correct at time of download.
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
In?ation and Mexican economic
growth: long-run relation and
threshold effects
W. Adria´n Risso and Edgar J. Sa´nchez Carrera
Department of Economics, University of Siena, Siena, Italy
Abstract
Purpose – The purpose of this paper is to estimate the long-run relationships and threshold effects
between in?ation and economic growth in Mexico.
Design/methodology/approach – The paper shows the existence of such relationship in a
cointegrated vector on economic growth (log of real gross domestic product (GDP)) and in?ation rate
?nding a corresponding elasticity signi?cantly negative. Moreover, the causal relationship between
these two series is studied using a more robust Granger causality test, without ?nding any directional
causality between them.
Findings – The estimated threshold model suggests 9 percent as the threshold level (i.e. structural
break point) of in?ation above which in?ation signi?cantly slows the Mexican economic growth.
Research limitations/implications – This paper uses the cointegration technique, and ?nds a
signi?cant and negative long-run relationship between in?ation and economic growth for the Mexican
economy. In addition, it is found that in?ation is weakly exogenous. In the period 1970-2007 real GDP
was elastic with respect to in?ation, and therefore, considering the estimated coef?cient, an increase of
1 percent on in?ation produces a decrease of 1.5 percent on real GDP. Since, for most of the period under
consideration Mexico experienced in?ation rates higher than 10 percent, this result is consistent with
most of the research suggesting that high levels of in?ation produce a negative effect on economic
growth.
Practical implications – The analysis could be useful for policymakers in providing some clue in
setting an optimal in?ation target. For instance, the Mexican Central Bank could apply an expansionary
monetary policy for supporting economic growth until the in?ation rate does not exceed the threshold
level. In fact, the threshold analysis suggests that if the in?ation rate exceeds 9 percent, then Mexico’s
current favorable economic performance might be jeopardized.
Originality/value – Speci?cally, this paper focuses on two questions: is there any long-run
relationship between economic growth and in?ation in Mexico? Is there a statistically signi?cant
threshold level of in?ation above which in?ation affects growth differently than at lower in?ation
rates in Mexico? Motivated by these questions, this present paper ?rst examines cointegration
techniques and then, threshold estimations.
Keywords Mexico, National economy, In?ation, Economic growth
Paper type Research paper
1. Introduction
The study of the effects of in?ation on economic growth continues to be an important
and complex topic in economics. If in?ation has real economic effects, then governments
can in?uence economic performance through monetary policy. Therefore, investigating
how in?ation affects economic growth pertains directly to the optimal design of
monetary policy. Results fromsuch studies are particularly important for economies like
Mexico because of its history of high in?ation and because its Central Bank has recently
adopted an in?ation targeting policy. Consider, for example the following two scenarios:
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1757-6385.htm
JFEP
1,3
246
Journal of Financial Economic Policy
Vol. 1 No. 3, 2009
pp. 246-263
qEmerald Group Publishing Limited
1757-6385
DOI 10.1108/17576380911041728
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
reducing in?ation byone percentage point when the rate is 20 percent whichresults in an
increase in the growth rate of 0.5 percent, compared to reducing in?ation by one
percentage point when the in?ation rate is around 5 percent, which results in a decrease
in the growth rate by 1 percent. If such scenarios hold, it could be more costly for a
low-in?ation country to decrease its in?ation by an additional point than it could be for a
country with a higher starting rate of in?ation (Andres and Hernando, 1999).
Furthermore, Mexico has recently been under pressure from international lending
agencies (the IMF, World Bank and ADB) to reduce its in?ation rate in order to boost
economic growth. But, as we will see, two extensive recent work by Bruno and Easterly
(1998) and Paul et al. (1997) that attempt to address this problem do not shed much light
on which approach would be an optimal policy response.
For these reasons, it is important to study whether there is a positive or negative
relationship between in?ation and economic growth and if this relationship differs for
low- and high-levels of in?ation. According to Grier and Grier (2006) there is little
theoretical consensus on how in?ation affects economic performance.
The topic has been widely studied and discussed by the structuralists and the
monetarists. The structuralists argue that in?ation is necessary for economic growth,
whereas the monetarists argue the opposite, that is, in?ation is detrimental to economic
growth (Mallik and Chowdhury, 2001). Such debate started in the 1950s, focused on
developing countries, which had long suffered from low-growth rates with high rates of
in?ation and larger de?cits in the balance of payments.
First, the monetarists argue that price stabilitypromotes economic growthandprotects
the balance of payments. They argue that in?ation is a major source of economic
instability because it weakens incentives for work and production, distorts the allocative
ef?ciency of the market mechanism, erodes international competitiveness of the domestic
industry, andreduces growth potential. The monetarists argue, furthermore, that in?ation
damages economic growth by lowering domestic and foreign savings, reducing ef?ciency
of resource allocation, and deteriorating the balance-of-payments. To monetarists, stable
prices are the starting point inthe process of economic development. The policy choice of a
country would be stabilization with growth, or stabilization without growth. Several
papers are typical of the monetarist tradition. Fischer and Modigliani (1980) suggested a
negative and nonlinear relationship between the rate of in?ation and economic growth
through the new growth theory mechanism. Cooley and Hansen (1991) proposed a model
where the agents decide the level of labor output, andanincrease inin?ationreduces labor
supply, and producing a decrease in economic production. Similarly, Barro (1995) pointed
out that the main effects of in?ation on economic performance are signi?cantly negative.
Second, and as an alternative, the structuralists argue that in?ation normally
accompanies economic growth in developing countries because structural rigidities and
bottlenecks insupplysectors prevent the elastic supplyof some basic commodities suchas
food, housing, energy, and transportation. Increased income as a result of growth would
expand demand for such basic commodities, and prices would rise. The structuralist
position is that economic dif?culties in developing countries have roots deeper than just
the results of in?ation. Thus, structuralists thought that in?ationary pressures and
deterioration in the balance of payments inevitably are attendant matters of economic
growth. In developing countries, there thus would be a trade-off relationship between
economic growth and in?ation and an attendant deterioration in balance of payments. If a
developing country wants stabilization of prices and balance of payments, it must reduce
In?ation and
Mexican
economic growth
247
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
the speed of economic growth, including a sacri?ce of employment. Among scholars who
support the structuralist’ position on a positive relationship between in?ation and
economic performance, Mundell (1965) and Tobin (1965) predict a positive relationship
between the rate of in?ation and the rate of capital accumulation, which in turn implies
a positive relationship to the rate of economic growth. They argue that, since money and
capital are substitutable, an increase in the rate of in?ation increases capital accumulation
by shifts in portfolios from money to capital and thereby stimulates a higher rate of
economic growth(see more details inDe Gregorio(1996)). Moreover, Fischer (1926) was the
?rst to establish a negative correlation between in?ation and unemployment, and Okun
(1962) found a negative correlation between unemployment and economic growth, then
frombothpropositions it canbe deduceda positive relationship betweeneconomic growth
and in?ation. Phillips (1958) proposed a positive relationship between in?ation and
unemployment implying the same type of relationship.
To sum-up, monetarists suggest that economic policy should emphasize stabilization
of aggregate demand via careful, stable monetary policy. However, monetarists do
not regard ?scal policies or other policy efforts to remove structural rigidities as
unimportant. The monetarists’ position is that though various policy efforts may
strengthenthe supplysector andpromote economic growth, the whole set of policyactions
should be taken within a stable monetary stance. Monetarists like economic policies that
rely upon the price mechanism rather than interventionist government policies. The
structuralists advocate more speci?c, micro-based policies to remove the bottlenecks
or rigidities inherent in the supply sectors. Because of the market disequilibria frequently
found in developing countries, the price mechanism does not function well enough
to convey appropriate signals regarding supply and demand conditions. As such,
government intervention in the allocation of resources is unavoidable so that overall
production capacity can be expanded more rapidly than when the price mechanismalone
plays a major role in resource allocation. In addition, monetary expansion may be
considered a necessary by-product of the growth policy pursuing structural changes.
In any case, the “structuralists vs monetarists” debate continues. For instance, Faria and
Carneiro (2001) investigate the relationship between in?ation and output in the context
of an economy facing persistent high in?ation and they ?nd that in?ation does not affect
real output in the long run, but that in the short-run in?ation negatively affects output.
Finally, scholars such as Sidrauski (1967) suggest that there is no relationship between
in?ation and economic growth, supporting the hypothesis of superneutrality of money.
Empirical studies arrive at various conclusions, from papers ?nding no relationship
among the variables to studies asserting the existence of a nonlinear relationship.
De Gregorio (1993) studies 12 Latin-American countries in the period 1950-1985 and
?nds a signi?cant and negative relationship between economic growth and in?ation.
Barro (1995) studies panel data from 100 countries during the period 1960-1990 and
suggest a negative relationship. However, Bruno and Easterly (1998) do not agree with
these results, asserting that the negative relationship is due to a bias in the sample, they
argue that Barro included countries with in?ation larger than 40 percent and these
outliers potentially dominated the results. Fischer (1993) asserts that in?ation, the ?scal
de?cit and exchange rate distortions negatively impact on economic growth. Andres
and Hernando (1999) analyze the OECDcountries in1960-1992 andthey ?nd that current
in?ation has never been correlated positively with income per capita in the long run,
and most of the time the relationship is negative for long periods of time.
JFEP
1,3
248
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
On the other hand, Sarel (1995) asserts that there is a nonlinear relationship between
in?ation and economic growth. Using 87 countries, he ?nds the existence of an in?ation
threshold of 8 percent. Above the threshold there is a negative relationship between
in?ation and economic growth, whereas under the threshold there is a positive but not
signi?cant relationship. Judson andOrphanides (1996) divide Sarel’s sample of countries
into three groups, and they ?nd similar results to Sarel, ?nding a threshold of 10 percent.
Ghosh and Phillips (1998a, b) study 145 countries in the period 1960-1990 again ?nding
similar results. Paul et al. (1997) study 70 countries (of which 48 are developing
economies) for the period 1960-1989. They ?nd no causal relationship between in?ation
and economic growth in 40 percent of the countries, bidirectional causality among
20 percent of the countries, and unidirectional causality for the rest (either in?ation to
growth or vice versa). The relationship was found to be negative in some cases and
positive in others. Burdekin et al. (2004), studying a similar sample, highlight that
nonlinearities behave quite differently for industrial economies than they appear to
behave for developing countries. They ?nd that the threshold at which in?ation ?rst
begins to seriously and negatively affect growth is around 8 percent for industrial
economies but around 3 percent for developing countries. Loungani and Sheets (1997)
study 26 transition economies during the period 1991-1994. In?ation was high in the
period and they ?nd a negative relationship between in?ation and economic growth.
Mexico, similar to the rest of Latin America, has a history of long periods of high
in?ation. According to Gagnon (2009) Mexico suffered three digit in?ation rates in the
late 1980s. In 1994 in?ation was successfully stabilized below10 percent. However, after
the collapse of the exchange rate in 1994, in?ation peaked at 92 percent in April 1995 and
real output per capita contracted by 9.5 percent in 1995. Mexico would wait until 1998 for
real gross domestic product (GDP) per capita to surpass its 1994 level and until 1999 for
in?ation to settle below 10 percent.
The Mexican case has been studied by several scholars. Mendoza (1998) ?nds
that in?ation has had no effect on Mexico’s long-run economic growth. Katz (2002)
asserts that in?ation is dangerous for Mexico’s economic growth, identifying a negative
correlation between the variables in Mexico during the period 1970-2002. Acevedo (2006)
applies the nonlinear approach proposed by Sarel (1995) to the Mexican economy.
He ?nds similar results to Sarel (1995) and Judson and Orphanides (1996), showing
evidence for a threshold of 8.1 percent for Mexico, where in?ation rates higher than this
level severely damage Mexico’s economic performance.
Grier and Grier (2006) present evidence on the real effects of in?ation and in?ation
uncertainty on Mexican output growth. Their main ?ndings are as follows:
.
in?ation uncertainty has a negative and signi?cant effect on growth;
.
once the effect of in?ation uncertainty is accounted for, lagged in?ation does not
have a direct negative effect on output growth; and
.
however, as predicted by Ball (1992), higher average in?ation raises in?ation
uncertainty, and the overall net effect of average in?ation on output growth in
Mexico is negative.
That is, the average in?ation rate is harmful to Mexican growth due to its impact on
in?ation uncertainty.
It seems, therefore, that there is no theoretical consensus on exactly how in?ation
affects economic performance. Much of the empirical literature searches for a negative
In?ation and
Mexican
economic growth
249
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
in?uence of in?ation on growth, yet many economic theories predict neutrality or even a
positive effect of average in?ation on economic performance. So, the aimof this paper is
to study the nature of the relationship between in?ation and economic growth and
speci?cally, this paper focuses on two questions[1]:
Q1. Is there any long-run relationship between economic growth and in?ation in
Mexico?
Q2. Is there a statistically signi?cant threshold level of in?ation above which
in?ation affects growth differently than at lower in?ation rates in Mexico?
Motivated by these questions, the paper ?rst examines cointegration techniques and
proceeds, second, to threshold estimations (Hansen, 1999, 2000). To study the threshold
level of in?ation, we adopt the econometric technique developed by Khan and Senhadji
(2001) analyzing a nonlinear or structural break effect that examines whether the impact
of in?ation on economic growth could be positive up to a certain threshold level, but
beyond this level the effect turns negative. The model provides appropriate procedures
for estimation and inference. Using the nonlinear least squares (NLS) estimation
technique, Khan and Senhadji (2001) estimated that the threshold levels for industrial
countries and developing countries were at 1-3 and 11-12 percent, respectively[2].
We found a threshold level of 9 percent of in?ation above which the effect of in?ation is
negative and statistically signi?cant. But, in?ation has a negative effect on long-run
growth only if the level of in?ation is above such a threshold level. We found a
statistically signi?cant long-run negative relationship between in?ation and economic
growth. Our results therefore corroborate the results produced by De Gregorio (1993),
Barro (1995) and Katz (2002).
The remainder of the paper is organized as follows. In Section 2, we introduce the data
set and the applied method. Section 3 presents the estimated long-run relationship using
the cointegration analysis. In Section 4, we estimate the in?ation threshold by applying
the technique suggested by Khan and Senhadji (2001). Finally, in Section 5, we draw
some concluding remarks.
2. Data set and methodology
We consider annual data of real GDP (Y), consumer price index (CPI) and investment
share (I) in 2000 constant prices (Mexican local currency) for the period 1970-2007.
In?ation (p) was computed as the growth rate of the CPI; both data on the growth rate of
real GDP and CPI were obtained from World Economic Outlook (WEO) database. Data
on the investment share of real GDP was obtained fromthe Penn World Tables 6.2. Data
on population (annual growth rate) (POP) was obtained fromOECDstatistics portal. We
applied a log transformation, since log transformations help, at least partially, to
eliminate the strong asymmetry in the distribution of in?ation (Sarel (1995) and Ghosh
and Phillips (1998a, b)).
We consider two econometric models to obtain the empirical results:
(1) In the ?rst part of the paper, we examine the short- and long-run relationships
between real GDP and p. First, we examine the rank of integration for the series
log (real GDP) and (p) using the augmented Dickey-Fuller (ADF) Dickey and
Fuller (1981) unit root and the Kwiatkowski-Phillips-Schmidt-Shin tests (KPSS).
Second, an unrestricted vector autoregressive (VAR) is estimated, which
JFEP
1,3
250
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
is known to be sensitive to the number of time lags (Banerjee et al., 1993).
To determine the appropriate number of time lags for the VAR, we employed the
Akaike information criterion (AIC) and the Schwartz Bayesian criterion. Third,
we examine the existence of a long-run relationship between GDP and p using a
vector error-correction model (VECM) after applying Johansen’s (1988, 1990, and
1995) cointegration technique. We conduct a test for weak exogeneity in order to
do inference. Finally, the causality between the variables is studied based on the
more robust Toda and Yamamoto (1995) Granger no-causality test which allows
the use of the Granger test in an integrated system.
(2) The second part of the paper attempts to ?nd the mechanism through which
in?ation nonlinearly affects long-run economic growth. Hence, we consider Khan
and Senhadji’s (2001) model to estimate a threshold level of in?ation above which
in?ation affects economic growth negatively.
3. Cointegration analysis
A spurious regression results when trending or unit root time series produce
non-stationary residuals, signi?cant ordinary least-squares (OLS) parameter estimates,
and a high-R
2
. Non-stationary residuals violate the standard assumptions that are
required to apply OLS methods. In this case, Phillips (1986) pointed out that
cointegration techniques must be applied. Recently, cointegration tests are widely used
in econometric practice. Cointegration tests suggest that restrictions be imposed on
multivariate or VAR models and can be used to test economic theories.
A?rst step in cointegration analysis is to study the stationarity of the series by using
unit root tests. Two popular techniques have been used: the ADFtest (Dickey and Fuller,
1981) and the KPSS test (Kwiatkowski, et al. 1992). These tests have been performed in
levels (i.e. log of real GDP and in?ation rate) as well as in ?rst differences. If the two time
series are integrated of the same order, then we can estimate a cointegrating regression.
Thus, we search for a long-run relationship between the two variables, followed by
applying a VECM to model the short-run dynamics. The model is represented in a
?rst-differenced error-correction form:
DY
t
¼ m þPY
t21
þ
X
i¼k21
i¼1
G
i
DY
t2i
þ1
i
z ð1Þ
where, Y ¼ (real GDP, in?ation) is a vector containing the variables and mis a vector of
constant terms. The matrix P conveys information about the long-run relationship
between the Y variables. The rank of P is the number of linearly independent and
stationary linear combinations of the variables.
McCallum (1984) asserts that incorrect signs can be produced if exogeneity is not
studied. To apply inference techniques, we must test weak exogeneity. Finally,
a modi?ed version of the Granger causality test is applied in order to analyze causality
between the variables.
3.1 Empirical results
A ?rst step in cointegration analysis is to study the stationarity of the series by using
unit root tests, such as the ADF and the KPSS tests. The null hypothesis of the KPSS test
is stationarity, complementing the ADF test. However, the ADF test has low power
against stationary near unit root processes. Nevertheless, Tables I and II show unit root
In?ation and
Mexican
economic growth
251
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
V
a
r
i
a
b
l
e
y
p
D
y
D
p
U
n
i
t
r
o
o
t
t
e
s
t
A
D
F
K
P
S
S
A
D
F
K
P
S
S
A
D
F
K
P
S
S
A
D
F
K
P
S
S
T
r
e
n
d
a
n
d
c
o
n
s
t
a
n
t
2
2
.
9
7
(
2
3
.
5
4
)
0
.
1
4
(
0
.
1
5
)
2
2
.
3
2
(
2
3
.
5
4
)
0
.
1
6
(
0
.
1
5
*
)
2
4
.
4
6
(
2
3
.
5
4
)
*
0
.
1
1
(
0
.
1
5
)
2
5
.
8
4
(
2
3
.
5
4
)
*
0
.
5
0
(
0
.
1
5
)
*
C
o
n
s
t
a
n
t
2
2
.
2
0
(
2
2
.
9
5
)
0
.
7
3
(
0
.
4
6
)
*
2
2
.
0
6
(
2
2
.
9
5
)
0
.
2
3
(
0
.
4
6
)
2
4
.
2
3
(
2
2
.
9
5
)
*
0
.
2
8
(
0
.
4
6
)
2
5
.
7
6
(
2
2
.
9
5
)
*
0
.
3
5
(
0
.
4
6
)
W
i
t
h
o
u
t
t
r
e
n
d
,
c
o
n
s
t
a
n
t
2
5
.
9
3
(
2
1
.
9
5
)
2
1
.
3
7
(
2
1
.
9
5
)
2
2
.
6
8
(
2
1
.
9
5
)
*
2
5
.
8
5
(
2
1
.
9
5
)
*
N
o
t
e
s
:
*
N
u
l
l
h
y
p
o
t
h
e
s
i
s
r
e
j
e
c
t
i
o
n
a
t
5
p
e
r
c
e
n
t
;
c
r
i
t
i
c
a
l
v
a
l
u
e
s
i
n
p
a
r
e
n
t
h
e
s
i
s
Table I.
ADF and KPSS unit root
tests
JFEP
1,3
252
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
tests for the variables in levels and in differences, the variables are expressed in
logarithmic form.
According to the tests, the time series are integrated processes of ?rst order, that is
they are I(1). Hence, we have to study the existence of a cointegrating relationship.
Banerjee et al. (1993) highlight that searching for a cointegration relation is equivalent to
searching for a statistical equilibriumbetween variables tending to growover time. The
discrepancy of this equilibrium can be modeled by a VECM (equation (1)). The VEC
model shows howthe variables return to the equilibriumafter suffering a shock. In order
to obtain the optimal VEC model, we applied the minimum AIC criterion. We estimate a
VAR with one lag (according to the minimum AIC) and test the cointegration
relationship. Table II indicates that one co-integrating relationship is found.
To do inference, we need to test the weak exogeneity of real GDP. The x
2
-statistic is
1.01 producing a p-value of 0.31. Therefore, we cannot reject the hypothesis that real
GDP is weakly exogenous at a 5 percent level of signi?cance.
In the second stage, the VEC model is employed to assess whether the economy
approaches equilibrium in the long run and to analyze the short-run dynamics of the
cointegrated time series variables. The VECis internally consistent if the two time series
variables are cointegrated of the same order or if they are stationary. Equation (2) shows
the long-run estimated relationship:
y ¼ þ7:02 21:56p ½231:53? ½3:01? ð2Þ
Equation (2) shows that the elasticity for in?ation-GDP was approximately 21.56
between 1970 and 2007. This elasticity means that an increase of the annual in?ation of
1 percent produced a decrease in GDP of more than 1.5 percent. This result coincides
with De Gregorio (1993), Barro (1995) and Katz (2002). We mentioned some work
suggesting that the negative relationship is found when in?ation is high. During most of
the period, Mexico suffered high in?ation and the cointegration relationship could
suggest a negative relationship. Since in?ation is weakly exogenous, we can perform
inference. We can say, therefore that high in?ation had a negative impact on economic
growth. Notice though that our result differs from the case shown by Mallik and
Chowdhury (2001). They examine the relationship between in?ation and GDP growth
for four South Asian countries (Bangladesh, India, Pakistan, and Sri Lanka). They ?nd
evidence of a long-run positive relationship between GDP growth rate and in?ation for
all four countries. Predominantly, these countries have not had high-in?ation crises
(except Bangladesh during 1972-1974 only); their in?ation rates of 7-10 percent can thus
be regarded as moderate.
Hypothesis Trace statistic CV at 0.05 p-value
None
a
21.44 20.26 0.034
At most 1 4.73 9.16 0.314
Hypothesis Max. Eig. Stat. CV at 0.05 p-value
None
a
16.71 15.89 0.037
At most 1 4.73 9.16 0.314
Note:
a
Indicates rejection of the null hypothesis at 5 percent
Source: Own calculations
Table II.
Johansen cointegration
test
In?ation and
Mexican
economic growth
253
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Though cointegration is required to examine the relationships, it does not indicate
the direction of the causal relationship. Granger (1988) proposed a test to study
causality, subsequently called Granger causality. But Granger causality does not imply
causality in a philosophical sense; rather it should be understood instead as a kind
of predetermination among variables. Dynamic Granger causality can be captured in the
VAR model. Notice that, the variables are integrated of the same order, and applying
the standard Granger causality test would be not appropriated. In this vein, Toda and
Yamamoto (1995) suggest an alternative procedure. When the variables are integrated,
they propose estimating a VAR model with (k þ d
max
) lags, where k is the standard
optimal number of lags and d
max
is the maximal order of integration that we suspect
might occur in the process. Once the VAR is estimated, we may test Granger causality
using only the ?rst k lags. For instance, if we consider the following equation from a
VAR model:
p
t
¼ g
0
þg
1
Y
t21
þg
2
Y
t22
þg
3
p
t21
þg
4
p
t22
ð3Þ
where, k ¼ 1 was selected according the minimum AIC and d
max
¼ 1, the null
hypothesis of non-causality from Y to p should be:
H
0
: g
1
¼ 0 ð4Þ
Equation (4) means:
H
0
real GDP does not Granger-cause in?ation
The hypothesis is tested using the Wald test. Toda and Yamamoto (1995) assert that
the Wald and LR tests are asymptotically equivalent in the present situation. Table III
shows the results for all the variables.
According to the Wald statistic, we cannot reject the null hypothesis of no causality.
Therefore, we do not ?nd evidence of Granger-causality among the variables.
These ?ndings have important policy implications. Even if we do not ?nd any
direction in Granger causality, the weak exogenous test indicates that high levels of
in?ation were harmful to economic growth in the period considered. Hence, our evidence
corresponds to the monetarist position, and caution is needed since periods of higher
in?ation may produce negative in?ation elasticities.
4. Threshold analysis
Khan and Senhadji (2001) calculate growth rates of macroeconomic variables using a log
transformation because a log transformation provides a best ?t in the class of nonlinear
models. Although the growth rates of all these variables are calculated using log
transformation method, a lot of volatility remains in the data (Figure 1).
Null hypothesis Wald-statistic p-value
Y does not Granger cause P 0.05 0.83
P does not Granger cause Y 0.08 0.78
Notes: We applied a VAR with k þ dmax ¼ 1 þ 1; p-values correspond to the x
2
distribution with
two degree of freedom
Source: Toda and Yamamoto (1995)
Table III.
Granger causality test
JFEP
1,3
254
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Therefore, in order to smooth out business cycle ?uctuations, the data set is smoothed
using the Hodrick-Prescott ?lter (Figure 2).
To test for the existence of a threshold effect, the following model was estimated:
y
t
¼ b
0
þb
1
logðp
t
Þ þb
2
d
p
*
t
½logðp
t
Þ 2logðp
*
Þ? þu
0
X
t
þ e
t
ð5Þ
d
p
*
t
¼
1 if p
t
. p
*
0 if p
t
# p
*
8
<
:
t ¼ 1; . . . ; T
where, y
t
is the growth rate of real GDP, p
t
is the in?ation rate based on the CPI, p
*
is the
thresholdlevel of in?ation, d
p
*
t
is a dummyvariable that takes a value of one for in?ation
levels greater than p
*
percent and zero otherwise, X
t
is a vector of control variables
which includes investment as a share of GDP and the population growth rate. Our study
Figure 1.
Log transformation
of series
12
8
4
0
–4
–8
1970 1975 1980 1985 1990 1995 2000 2005
LNINF GDP
Figure 2.
Series smoothed
50
40
30
20
10
0
1970 1975 1980 1985 1990 1995 2000 2005
GDP filter
Invest filter
Infilter
Pop filter
In?ation and
Mexican
economic growth
255
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
is based on a four-variable model consisting of economic growth, in?ation, population,
and investment[3]. The parameter p
*
represents the threshold in?ation level with the
property that the relationship between output growth and in?ation is given by:
.
low-in?ation: b
1
; and
.
high in?ation: b
1
þ b
2
.
High in?ation means that when the long-run in?ation estimate is signi?cant then b
1
and
b
2
would be summed to see their impact ongrowth and that would produce the threshold
level of in?ation. By estimating regressions for different values of p
*
, chosen in
ascending order (i.e. 1, 2 and so on), the optimal value of p
*
is obtained by ?nding the
value that minimizes the residual sumof squares (RSS)[4]. In?ation beyond this level has
a signi?cant and negative impact on economic growth. This procedure has become
widely accepted in the literature. However, this process is tedious since it requires
estimating the equation several times for different values of p
*
. Nevertheless, statistical
rigor requires that it be done.
4.1 Empirical results
By using NLS to estimate equation (5), we can obtain the exact value of the threshold
in?ation level. The minimum RSS of each regression considers p
*
¼ 1 to p
*
¼ 13.
However, the values of p
*
when RSS is minimized ranges from 7-11 percent. The
estimated results are reported in Table IV.
According to the p-values, the estimator parameters are statistically signi?cant at very
lowvalues. Therefore, in the estimation process the threshold level of in?ation is observed
at the 9 percent level where RSS is minimized. Hence, while in?ation belowthis threshold
level has no signi?cant effect on economic growth (i.e. statistically insigni?cant at
5 percent level), in?ation rates above the threshold have a signi?cant and negative effect
on economic growth.
Fromthe estimated results, at lower threshold in?ation levels (p
*
less than 9 percent)
there is no statistically signi?cant relationship between the dummy of threshold level of
in?ation and economic growth. As p
*
is greater than9 percent, a statisticallysigni?cant
relationship is observed between economic growth and the dummy of threshold level
of in?ation which continues upto an11-percent in?ation rate, whichmeans that in?ation
has a negative statistically signi?cant effect on Mexican economic growth for in?ation
levels greater than 9 percent.
Since 1999, the Banco de Me´xico announced a mediumtermin?ation objective, and since
2000 the Central Bank publishes quarterly in?ation reports to monitor the in?ationary
process, to analyze in?ationprospects andto discuss the conduct of monetarypolicyandthe
balance of risks for future in?ation (BANXICO, 2008). Currently, Mexico is considered to
have in place the main components of an in?ation targeting framework: an independent
monetary authority (since 1993) that has in?ation as its only policy objective, a ?exible
exchange rate regime, the absence of other nominal anchors and a “transparent” framework
for the implementation of monetary policy (Schmidt-Hebbel and Werner, 2002).
In this vein, our results follow from the in?ation targeting aims that The Mexican
Central Bank de?ned in a series of annual in?ation targets since 1999. The target ceiling
started at 13 percent for 1999, decreased to 10 percent by 2000, and was followed
by additional decreases to 9 percent for the 2001 and 2002, 4.05 percent in 2006, and
3.76 percent in 2007.
JFEP
1,3
256
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
4.1.1 2SLS analysis. In equation (5), in?ation may not be exogenous and consequently
the coef?cient estimates may be biased. As Fischer (1993) pointed out, the causality is
more likely to be unidirectional from in?ation to growth, in which case the problem of
simultaneity bias may not be very important. Consequently, a Granger Causality test is
applied to measure the linear causation between in?ation and economic growth. Table V
shows that the null hypothesis is not rejected, which means that GDP (economic growth)
Granger causes in?ation. The causality between two variables is unidirectional since the
null hypothesis that in?ation causes output growth is not rejected at a 5 percent level of
signi?cance, which supports the hypothesis that there is no feedback from output
growth to in?ation.
Hence, in?ation causes economic growth (in the sense of either a positive sign or
negative sign on its coef?cient) but economic growth is not only explained by in?ation
rate, then the model could be incorrectly speci?ed, implying a speci?cation bias in the
Null hypothesis F-statistic p-value
In?ation does not Granger cause economic growth 1.50237 0.23837
Economic growth does not Granger cause in?ation 6.44918 0.00455
*
Note:
*
Indicates rejection of the null hypothesis at 5 percent
Table V.
Pairwise Granger
causality test
p
*
Variable Coef?cient SE t-statistic Prob. RSS
C 226.57851 2.192942 212.12002 0.0000
Log(p) 20.488365 0.073403 26.653202 0.0000
7 percent Break7 20.092338 0.031913 22.893442 0.0067 10.40
Pop 1.503096 0.064185 23.41810 0.0000
Inv 9.853121 0.742647 13.26758 0.0000
C 226.40593 2.110052 212.51435 0.0000
Log(p) 20.545863 0.079201 26.892144 0.0000
8 percent Break8 20.090380 0.026949 23.353742 0.0020 1.78
Pop 1.521239 0.062812 24.21894 0.0000
Inv 9.822111 0.711538 13.80405 0.0000
C 226.74910 1.963577 213.62264 0.0000
Log(p) 20.569815 0.075737 27.523557 0.0000
9 percent Break9 20.086855 0.021997 23.948469 0.0004 1.62
Pop 1.494305 0.058858 25.38810 0.0000
Inv 9.970370 0.660360 15.09839 0.0000
C 226.81594 1.969623 213.61476 0.0000
Log(p) 20.574085 0.077255 27.431035 0.0000
10 percent Break10 20.075255 0.019342 23.890821 0.0005 1.64
Pop 1.493487 0.059123 25.26058 0.0000
Inv 9.996084 0.662225 15.09470 0.0000
C 227.70299 2.028275 213.65840 0.0000
Log(p) 20.553976 0.082645 26.703076 0.0000
11 percent Break11 20.058091 0.017901 23.245056 0.0027 1.81
Pop 1.496378 0.062279 24.02695 0.0000
Inv 10.30664 0.682291 15.10593 0.0000
Note: Where break number is de?ned by ½logðp
t
Þ 2logðp
*
Þ?
Table IV.
Test results for the
threshold effects
estimation
In?ation and
Mexican
economic growth
257
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
estimated model (5), that is, exclusion of other relevant variables for a growth equation.
To check this bias, in?ation, real GDP growth, investment, and population growth are
used as instruments[5]. In order to control for endogeneity, at least partially, the model is
now estimated using two-stage least squares (2SLS), where the set of instruments
includes the lag of in?ation, the lag of real GDP, the lag of investment and the lag of the
population rate[6]. The results of 2SLS also suggest a 9 percent threshold in?ation level
(Table VI).
Both estimated models (Table IV (NLS) and Table VI (2SLS)) provide identical
threshold levels of in?ation and the values of estimated coef?cients also remain close
in both models. Both results indicate a 9 percent threshold in?ation level for economic
growth in Mexico. This threshold level lies within the range found by Khan and Senhadji
(2001), and it can be interpreted as the threshold level for a transition economy. These
results are similar to Sarel (1995), Judson andOrphanides (1996) in the international case,
and Acevedo (2006) who ?nds a threshold of 8.1 percent for Mexico, where in?ation rates
greater than this level would severely damage Mexico’s economic performance.
5. Concluding remarks
The main objective of this paper was to examine whether a relationship exists between
economic growth and in?ation and, if so, its nature. Even if there is neither a theoretical
p
*
Variable Coef?cient SE t-statistic Prob. RSS
C 227.60274 1.065922 22.589566 0.09930556
Log(p) 20.719575 0.358756 22.005752 0.37083333
7 percent Break7 20.677079 0.460916 21.468986 1.05277778 25.14
Pop 1.223486 0.452033 2.706632 0.075
Inv 9.976084 3.516970 2.836557 0.0078
C 220.07569 4.742451 24.233189 0.0002
Log(p) 21.041340 0.211722 24.918432 0.0000
8 percent Break8 20.317888 0.099192 23.204783 0.0031 5.18
Pop 1.681314 0.151782 1.107718 0.0000
Inv 7.704655 1.588038 4.851682 0.0000
C 220.76328 3.717135 25.585830 0.0000
Log(p) 21.020460 0.163859 26.227659 0.0000
9 percent Break9 20.225560 0.055577 24.058510 0.0003 3.29
Pop 1.630786 0.120352 13.55017 0.0000
Inv 8.031724 1.241283 6.470503 0.0000
C 220.63593 3.729246 25.533540 0.0000
Log(p) 21.045673 0.168227 26.215838 0.0000
10 percent Break10 20.196727 0.048571 24.050300 0.0003 3.31
Pop 1.636986 0.120465 13.58889 0.0000
Inv 7.999773 1.244817 6.426465 0.0000
C 218.82909 3.941537 24.777092 0.0000
Log(p) 21.194979 0.200614 25.956612 0.0000
11 percent Break11 20.166932 0.042509 23.927015 0.0004 3.52
Pop 1.770391 0.126191 14.02943 0.0000
Inv 7.461593 1.309084 5.699856 0.0000
Note: Where break number is de?ned by ½logðp
t
Þ 2logðp
*
Þ?
Table VI.
Test results for
the threshold effects
of 2SLS estimation
JFEP
1,3
258
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
nor an empirical consensus about the relationships between these variables, some studies
seem to suggest that an in?ation rate threshold exists above which the relationship
between in?ation and economic growth is negative and below which the relationship
is positive or zero. This would mean that high levels of in?ation may harm the economy.
However, setting zero in?ation could also be harmful when production is growing.
According to Akerlof et al. (1996) ?rms cannot support shocks of demand under zero
in?ation. These shocks could then produce inef?cient, low levels of employment. This
suggests that there may be a minimum ef?cient level of in?ation. Furthermore, Sinn and
Reuter (2001) assert that estimating a minimum level of in?ation is an important step in
in?ation targeting to prevent de?ation.
We used the cointegration technique, and found a signi?cant and negative long-run
relationship between in?ation and economic growth for the Mexican economy.
In addition, we found that in?ation is weakly exogenous. In the period 1970-2007 real
GDP was elastic with respect to in?ation, and therefore, considering the estimated
coef?cient, an increase of 1 percent on in?ation produces a decrease of 1.5 percent on
real GDP. Since, for most of the period under consideration Mexico experienced
in?ation rates higher than 10 percent, this result is consistent with most of the research
suggesting that high levels of in?ation produce a negative effect on economic growth.
Using a more robust Granger causality test, we found no predetermination between the
variables. However, the weak exogeneity of in?ation allows us to do inference and
shows that during the period 1970-2007 increasing (decreasing) in?ation was harmful
(bene?cial) to the economy.
As mentioned before, Mexico has an in?ation targeting policy. Thus, our second
analysis could be useful for policymakers in providing some clue in setting an optimal
in?ation target. For instance, the Mexican Central Bank could apply an expansionary
monetary policy for supporting economic growth until the in?ation rate does not
exceed the threshold level. In fact, our threshold analysis suggests that if the in?ation
rate exceeds 9 percent, then Mexico’s current favorable economic performance might
be jeopardized. Mishkin (2000) asserts that one of the advantages of the in?ation
targets is to respond to domestic shocks. International lending agencies gave policy
advice to Mexico, outlining how it should attempt to reduce in?ation to a very low level
to get positive effects on economic growth. The relevant question becomes what should
this level of low-in?ation be?
To conclude, as is now well-known, the Mexican economy has been devastated by
the current economic crisis. Mexico has experienced in?ation in excess of 5 percent
because of increasing costs for food and for fuel, with commensurate decreases in
demand for basic commodities affecting GDP. But, though the in?ation target is now
3 percent, Mexico is experiencing in?ation around 5 percent (4.89 percent at most recent
estimates). As a consequence, of the crisis, Mexico is unlikely to attain in?ation of
3 percent until later in 2010, when optimistic forecasts show that its growth may also
return to 3 percent or higher. Though debates about in?ation targeting, Mexico will
continue to feel the brunt of the economic crisis in conjunction with the other problems
endemic to its economy. That is, according to our results and the empirical evidence in
Mexico, the in?ation targeting is far from being a serious problem to attain a sustained
economic growth in the country. Nowadays, other problems like corruption, drug
traf?cking, and unemployment are crucial barriers for the Mexican growing.
In?ation and
Mexican
economic growth
259
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Notes
1. We chose this focus since the Banco de Mexico (or Mexican Central Bank) is required by the
Constitution (Article 28, paragraph 6) to achieve price stability as its primary objective.
According to the bank’s 1994 annual report, in?ation abatement had been given special
emphasis in Mexican economic policy since the end of 1987. Reducing in?ation was viewed as
a necessary condition for attaining social justice and long-run economic growth (Banco de
Me´xico 1995, p. 50).
2. Khan and Senhadji (2001) used an unbalanced panel with 140 countries over 40 years to
precisely examine the nonlinear relationship between in?ation and growth. The estimate of
the threshold level was 1-3 percent for industrial countries and 11-12 percent for developing
countries.
3. Population growth rate and investment are used as control variables. The reason for choosing
these variables is their authenticity in empirical literature on growth. Solow (1956) and Swan
(1956) who developed the ?rst neo-classical models of growth, take the rate of growth of
population as one of exogenous variables in their model to show that the faster the rate of
population growth, the poorer the country. Fischer (1993) includes investment in his model to
show that in?ation reduces growth by reducing investment and productivity growth.
Moreover, Mankiwet al. (1992) also include investment growth and population growth in their
growth model.
4. As KhanandSenhadji (2001) pointedout testing for no thresholdeffects means H
0
: b
2
¼ 0, but
under the null hypothesis, the threshold p
*
is not identi?ed, so classical tests, such as the
t-test, have nonstandard distributions. They suggest following the Hansen (1996, 1999)
methodology to simulate an asymptotic distribution of the likelihood ratio test of H
0
, which
implies to minimize the RSS.
5. For instance, investment (as share of GDP) is also likely to be endogenous to growth.
6. These are valid instruments only if the error term in equation (5) is not autocorrelated.
The DW and Breusch-Godfrey tests reject autocorrelation at 5 percent for all estimations.
References
Acevedo, E. (2006), “In?acio´n y crecimiento econo´mico en Me´xico: una relacio´n no lineal”,
Econom? ´a Mexicana Nueva E
´
poca, Vol. 15 No. 2, pp. 199-249.
Akerlof, G., Dickens, W. and Parry, G. (1996), “The macroeconomics of low in?ation”, Brookings
Papers on Economic Activity, Vol. 27 No. 1, pp. 1-76.
Andres, J. and Hernando, I. (1999), “Does in?ation harm economic growth? Evidence for the
OECD countries”, in Feldstein, M. (Ed.), The Costs and Effects of Price Stability,
The University of Chicago Press, Chicago, IL, pp. 315-41.
Ball, L. (1992), “Why does high in?ation raise in?ation uncertainty?”, Journal of Monetary
Economics, Vol. 29 No. 3, pp. 371-88.
Banco de Me´xico (1995), Informe Annual: 1994, Banco de Me´xico, Me´xico City.
Banerjee, A., Dolado, J., Galbraith, J. and Hendry, D. (1993), Co-integration, Error Correction, and
the Econometric Analysis of the Non-Stationary Data, Oxford University Press, Oxford.
BANXICO (2008), Quarterly In?ation Reports, The Central Bank of Mexico, BANXICO, Mexico
City, available: www.banxico.org.mx/publicaciones/JSP/b_in?ationreport.jsp
Barro, R.J (1995), “In?ation and economic growth”, NBER Working Papers No. 5326,
National Bureau of Economic Research, Cambridge, MA.
Bruno, M. and Easterly, W. (1998), “In?ation crises and long-run growth”, Journal of Monetary
Economics, Vol. 41 No. 1, pp. 3-26.
JFEP
1,3
260
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Burdekin, R., Denzau, A., Keil, M., Sitthiyot, T. and Willett, T. (2004), “When does in?ation hurt
economic growth? Different nonlinearities for different economies”, Journal of
Macroeconomics, Vol. 26 No. 3, pp. 519-32.
Cooley, T. and Hansen, G. (1991), “The in?ation tax in a real business cycle model”, American
Economic Review, Vol. 79 No. 4, pp. 733-48.
De Gregorio, J. (1993), “In?ation, taxation and long-run growth”, Journal of Monetary Economics,
Vol. 31 No. 3, pp. 271-98.
De Gregorio, J. (1996), “In?ation, growth, and central banks: theory and evidence”, Policy
Research Working Paper Series No. 1575, The World Bank, Washington, DC.
Dickey, D.A. and Fuller, W.A. (1981), “Likelihood ratio statistics for autoregressive time series
with a unit root”, Econometrica, Vol. 49 No. 4, pp. 1057-72.
Faria, R.J. and Carneiro, F. (2001), “Does high in?ation affect growth in the long and short run?”,
Journal of Applied Economics, Vol. 4 No. 1, pp. 89-105.
Fischer, I. (1926), “A statistical relationship between unemployment and price changes”,
International Labor Review, Vol. 13.
Fischer, S. (1993), “The role of macroeconomic factors in economic growth”, Journal of Monetary
Economics, Vol. 32 No. 2, pp. 485-512.
Fischer, S. and Modigliani, F. (1980), “Towards and understanding of the real effects and costs of
in?ation”, NBER Working Papers No. 0303, National Bureau of Economic Research,
Cambridge, MA.
Gagnon, E. (2009), “Price setting during low and high in?ation: evidence from Mexico”,
The Quarterly Journal of Economics, Vol. 124 No. 3, pp. 1221-63.
Ghosh, A. and Phillips, S. (1998a), “In?ation, disin?ation, and growth”, IMF Working Paper
No.WP/98/68, IMF, Washington, DC.
Ghosh, A. and Phillips, S. (1998b), “Warning: in?ation may be harmful to your growth”, IMF
Staff Papers, Vol. 45, IMF, Washington, DC.
Granger, C. (1988), “Some recent developments in a concept of causality”, Journal of
Econometrics, Vol. 39 Nos 1/2, pp. 199-211.
Grier, K. and Grier, R. (2006), “On the real effects of in?ation and in?ation uncertainty in Mexico”,
Journal of Development Economics, Vol. 80 No. 2, pp. 478-500.
Hansen, B. (1996), “Inference when a nuisance parameter is not identi?ed under the null
hypothesis”, Econometrica, Vol. 64 No. 2, pp. 413-30.
Hansen, B. (1999), “Threshold effects in non-dynamic panels: estimation, testing, and inference”,
Journal of Econometrics, Vol. 93 No. 2, pp. 345-68.
Hansen, B. (2000), “Sample splitting and threshold estimation”, Econometrica, Vol. 68
No. 3, pp. 575-604.
Johansen, S. (1988), “Statistical analysis of cointegration vectors”, Journal of Economic Dynamics
and Control, Vol. 12 Nos 2/3, pp. 231-54.
Johansen, S. (1995), Likelihood-based Inference in Cointegrated Vector Autorregressive Models,
Oxford University Press, Oxford.
Johansen, S. and Juselius, K. (1990), “Maximum likelihood estimation and inference on
cointegration with applications to the demand for money”, Oxford Bulletin of Economics
and Statistics, Vol. 52 No. 2, pp. 169-210.
Judson, R. and Orphanides, A. (1996), “In?ation, volatility and growth, ?nance and economics”,
Discussion Paper No. 96-19, Board of Governors, Federal Reserve of United States,
Washington, DC.
In?ation and
Mexican
economic growth
261
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Katz, I. (2002), “In?acio´n, crecimiento, pobreza y desigualdad en Me´xico”, en La in?acio´n
en Me´xico, Gaceta de Econom? ´a, Instituto Tecnolo´gico Auto´nomo de Me´xico, Mexico City.
Khan, M.S. and Senhadji, A.S. (2001), “Threshold effects in the relationship between in?ation and
growth”, IMF Staff Papers, Vol. 48 No. 1.
Kwiatkowski, D., Phillips, P., Schmidt, P. and Shin, Y. (1992), “Testing the null hypothesis
of stationarity against the alternative of a unit root”, Journal of Econometrics, Vol. 54
Nos 1-3, pp. 159-78.
Loungani, P. and Sheets, N. (1997), “Central bank independence, in?ation, and growth in
transition economies”, Journal of Money, Credit & Banking, Vol. 29 No. 3, pp. 381-99.
McCallum, B. (1984), “On low-frequency estimates of long-run relationships in macroeconomics”,
Journal of Monetary Economics, Vol. 14 No. 1, pp. 3-14.
Mallik, G. and Chowdhury, A. (2001), “In?ation and economic growth: evidence from South
Asian countries”, Asian Paci?c Development Journal, Vol. 8 No. 1, pp. 123-35.
Mankiw, N., Romer, D. and Weil, D. (1992), “A contribution to the empirics of economic growth”,
The Quarterly Journal of Economics, Vol. 107 No. 2, pp. 407-37.
Mendoza, M. (1998), “In?acio´n y crecimiento econo´mico en Me´xico, Monetaria”, revista
del CEMLA, Vol. 21 No. 2.
Mishkin, F. (2000), “In?ation targeting in emerging market countries”, The American Economic
Review, Vol. 90 No. 2, pp. 105-9.
Mundell, R. (1965), “Growth, stability and in?ationary ?nance”, Journal of Political Economy,
Vol. 73 No. 2, pp. 97-109.
Okun, A. (1962), “Potential GNP: its measurement and signi?cance”, in Pechman, J. (Ed.),
Economics for Policymaking, MIT Press, Cambridge, MA, 1983.
Paul, S., Kearney, C. and Chowdhury, K. (1997), “In?ation and economic growth: a multi-country
empirical analysis”, Applied Economics, Vol. 29 No. 10, pp. 1387-401.
Phillips, A. (1958), “The relationship between unemployment and the rate of change of money
wages in the United Kingdom, 1861-1957”, Economica, Vol. 25 No. 100, pp. 283-99.
Phillips, P. (1986), “Understanding spurious regressions in econometrics”, Journal of
Econometrics, Vol. 33 No. 3, pp. 311-40.
Sarel, M. (1995), “Nonlinear effects of in?ation on economic growth”, IMF Working Paper
Nos 95/56, International Monetary Fund, Washington, DC.
Schmidt-Hebbel, K. and Werner, A. (2002), “In?ation targeting in Brazil Chile and Mexico:
performance, credibility and exchange rate”, Central Bank of Chile Working Paper No. 171,
Central Bank of Chile, Santiago.
Sidrauski, M. (1967), “In?ation and economic growth”, Journal of Political Economy, Vol. 75 No. 6,
pp. 796-810.
Sinn, H. and Reuter, M. (2001), “The minimum in?ation rate for Euroland”, NBER Working
Paper, 8085, National Bureau of Economic Research, Cambridge, MA.
Solow, R. (1956), “A contribution to the theory of economic growth”, The Quarterly Journal of
Economics, Vol. 70 No. 1, pp. 65-94.
Swan, T. (1956), “Economic growth and capital accumulation”, Economic Record, Vol. 32
No. 63, pp. 344-61.
Tobin, J. (1965), “Money and economic growth”, Econometrica, Vol. 33 No. 4, pp. 671-84.
Toda, H. and Yamamoto, T. (1995), “Statistical inference in vector autorregressions with possibly
integrated processes”, Journal of Econometrics, Vol. 66 Nos 1/2, pp. 225-50.
JFEP
1,3
262
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Further reading
Engle, R. and Granger, C. (1987), “Co-integration and error correction: representation, estimation
and testing”, Econometrica, Vol. 55 No. 2, pp. 251-76.
Sweidan, O.D. (2004), “Does in?ation harm economic growth in Jordan? An econometric analysis
for the period 1970-2000”, International Journal of Applied Econometrics and Quantitative
Studies, Vol. 1 No. 2, pp. 41-66.
Corresponding author
Edgar J. Sa´nchez Carrera can be contacted at: [email protected]
In?ation and
Mexican
economic growth
263
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
doc_392831487.pdf
The purpose of this paper is to estimate the long-run relationships and threshold effects
between inflation and economic growth in Mexico.
Journal of Financial Economic Policy
Inflation and Mexican economic growth: long-run relation and threshold effects
W. Adrián Risso Edgar J . Sánchez Carrera
Article information:
To cite this document:
W. Adrián Risso Edgar J . Sánchez Carrera, (2009),"Inflation and Mexican economic growth: long-run
relation and threshold effects", J ournal of Financial Economic Policy, Vol. 1 Iss 3 pp. 246 - 263
Permanent link to this document:
http://dx.doi.org/10.1108/17576380911041728
Downloaded on: 24 January 2016, At: 21:36 (PT)
References: this document contains references to 54 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 1334 times since 2009*
Users who downloaded this article also downloaded:
Ahmad Zubaidi Baharumshah, Siew-Voon Soon, (2014),"Inflation, inflation uncertainty and output growth:
what does the data say for Malaysia?", J ournal of Economic Studies, Vol. 41 Iss 3 pp. 370-386 http://
dx.doi.org/10.1108/J ES-05-2012-0073
Girijasankar Mallik, Anis Chowdhury, (2011),"Effect of inflation uncertainty, output uncertainty and oil
price on inflation and growth in Australia", J ournal of Economic Studies, Vol. 38 Iss 4 pp. 414-429 http://
dx.doi.org/10.1108/01443581111160879
Ming-Yu Cheng, Hui-Boon Tan, (2002),"Inflation in Malaysia", International J ournal of Social Economics,
Vol. 29 Iss 5 pp. 411-425 http://dx.doi.org/10.1108/03068290210423532
Access to this document was granted through an Emerald subscription provided by emerald-srm:115632 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for
Authors service information about how to choose which publication to write for and submission guidelines
are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as
providing an extensive range of online products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee
on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.
*Related content and download information correct at time of download.
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
In?ation and Mexican economic
growth: long-run relation and
threshold effects
W. Adria´n Risso and Edgar J. Sa´nchez Carrera
Department of Economics, University of Siena, Siena, Italy
Abstract
Purpose – The purpose of this paper is to estimate the long-run relationships and threshold effects
between in?ation and economic growth in Mexico.
Design/methodology/approach – The paper shows the existence of such relationship in a
cointegrated vector on economic growth (log of real gross domestic product (GDP)) and in?ation rate
?nding a corresponding elasticity signi?cantly negative. Moreover, the causal relationship between
these two series is studied using a more robust Granger causality test, without ?nding any directional
causality between them.
Findings – The estimated threshold model suggests 9 percent as the threshold level (i.e. structural
break point) of in?ation above which in?ation signi?cantly slows the Mexican economic growth.
Research limitations/implications – This paper uses the cointegration technique, and ?nds a
signi?cant and negative long-run relationship between in?ation and economic growth for the Mexican
economy. In addition, it is found that in?ation is weakly exogenous. In the period 1970-2007 real GDP
was elastic with respect to in?ation, and therefore, considering the estimated coef?cient, an increase of
1 percent on in?ation produces a decrease of 1.5 percent on real GDP. Since, for most of the period under
consideration Mexico experienced in?ation rates higher than 10 percent, this result is consistent with
most of the research suggesting that high levels of in?ation produce a negative effect on economic
growth.
Practical implications – The analysis could be useful for policymakers in providing some clue in
setting an optimal in?ation target. For instance, the Mexican Central Bank could apply an expansionary
monetary policy for supporting economic growth until the in?ation rate does not exceed the threshold
level. In fact, the threshold analysis suggests that if the in?ation rate exceeds 9 percent, then Mexico’s
current favorable economic performance might be jeopardized.
Originality/value – Speci?cally, this paper focuses on two questions: is there any long-run
relationship between economic growth and in?ation in Mexico? Is there a statistically signi?cant
threshold level of in?ation above which in?ation affects growth differently than at lower in?ation
rates in Mexico? Motivated by these questions, this present paper ?rst examines cointegration
techniques and then, threshold estimations.
Keywords Mexico, National economy, In?ation, Economic growth
Paper type Research paper
1. Introduction
The study of the effects of in?ation on economic growth continues to be an important
and complex topic in economics. If in?ation has real economic effects, then governments
can in?uence economic performance through monetary policy. Therefore, investigating
how in?ation affects economic growth pertains directly to the optimal design of
monetary policy. Results fromsuch studies are particularly important for economies like
Mexico because of its history of high in?ation and because its Central Bank has recently
adopted an in?ation targeting policy. Consider, for example the following two scenarios:
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1757-6385.htm
JFEP
1,3
246
Journal of Financial Economic Policy
Vol. 1 No. 3, 2009
pp. 246-263
qEmerald Group Publishing Limited
1757-6385
DOI 10.1108/17576380911041728
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
reducing in?ation byone percentage point when the rate is 20 percent whichresults in an
increase in the growth rate of 0.5 percent, compared to reducing in?ation by one
percentage point when the in?ation rate is around 5 percent, which results in a decrease
in the growth rate by 1 percent. If such scenarios hold, it could be more costly for a
low-in?ation country to decrease its in?ation by an additional point than it could be for a
country with a higher starting rate of in?ation (Andres and Hernando, 1999).
Furthermore, Mexico has recently been under pressure from international lending
agencies (the IMF, World Bank and ADB) to reduce its in?ation rate in order to boost
economic growth. But, as we will see, two extensive recent work by Bruno and Easterly
(1998) and Paul et al. (1997) that attempt to address this problem do not shed much light
on which approach would be an optimal policy response.
For these reasons, it is important to study whether there is a positive or negative
relationship between in?ation and economic growth and if this relationship differs for
low- and high-levels of in?ation. According to Grier and Grier (2006) there is little
theoretical consensus on how in?ation affects economic performance.
The topic has been widely studied and discussed by the structuralists and the
monetarists. The structuralists argue that in?ation is necessary for economic growth,
whereas the monetarists argue the opposite, that is, in?ation is detrimental to economic
growth (Mallik and Chowdhury, 2001). Such debate started in the 1950s, focused on
developing countries, which had long suffered from low-growth rates with high rates of
in?ation and larger de?cits in the balance of payments.
First, the monetarists argue that price stabilitypromotes economic growthandprotects
the balance of payments. They argue that in?ation is a major source of economic
instability because it weakens incentives for work and production, distorts the allocative
ef?ciency of the market mechanism, erodes international competitiveness of the domestic
industry, andreduces growth potential. The monetarists argue, furthermore, that in?ation
damages economic growth by lowering domestic and foreign savings, reducing ef?ciency
of resource allocation, and deteriorating the balance-of-payments. To monetarists, stable
prices are the starting point inthe process of economic development. The policy choice of a
country would be stabilization with growth, or stabilization without growth. Several
papers are typical of the monetarist tradition. Fischer and Modigliani (1980) suggested a
negative and nonlinear relationship between the rate of in?ation and economic growth
through the new growth theory mechanism. Cooley and Hansen (1991) proposed a model
where the agents decide the level of labor output, andanincrease inin?ationreduces labor
supply, and producing a decrease in economic production. Similarly, Barro (1995) pointed
out that the main effects of in?ation on economic performance are signi?cantly negative.
Second, and as an alternative, the structuralists argue that in?ation normally
accompanies economic growth in developing countries because structural rigidities and
bottlenecks insupplysectors prevent the elastic supplyof some basic commodities suchas
food, housing, energy, and transportation. Increased income as a result of growth would
expand demand for such basic commodities, and prices would rise. The structuralist
position is that economic dif?culties in developing countries have roots deeper than just
the results of in?ation. Thus, structuralists thought that in?ationary pressures and
deterioration in the balance of payments inevitably are attendant matters of economic
growth. In developing countries, there thus would be a trade-off relationship between
economic growth and in?ation and an attendant deterioration in balance of payments. If a
developing country wants stabilization of prices and balance of payments, it must reduce
In?ation and
Mexican
economic growth
247
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
the speed of economic growth, including a sacri?ce of employment. Among scholars who
support the structuralist’ position on a positive relationship between in?ation and
economic performance, Mundell (1965) and Tobin (1965) predict a positive relationship
between the rate of in?ation and the rate of capital accumulation, which in turn implies
a positive relationship to the rate of economic growth. They argue that, since money and
capital are substitutable, an increase in the rate of in?ation increases capital accumulation
by shifts in portfolios from money to capital and thereby stimulates a higher rate of
economic growth(see more details inDe Gregorio(1996)). Moreover, Fischer (1926) was the
?rst to establish a negative correlation between in?ation and unemployment, and Okun
(1962) found a negative correlation between unemployment and economic growth, then
frombothpropositions it canbe deduceda positive relationship betweeneconomic growth
and in?ation. Phillips (1958) proposed a positive relationship between in?ation and
unemployment implying the same type of relationship.
To sum-up, monetarists suggest that economic policy should emphasize stabilization
of aggregate demand via careful, stable monetary policy. However, monetarists do
not regard ?scal policies or other policy efforts to remove structural rigidities as
unimportant. The monetarists’ position is that though various policy efforts may
strengthenthe supplysector andpromote economic growth, the whole set of policyactions
should be taken within a stable monetary stance. Monetarists like economic policies that
rely upon the price mechanism rather than interventionist government policies. The
structuralists advocate more speci?c, micro-based policies to remove the bottlenecks
or rigidities inherent in the supply sectors. Because of the market disequilibria frequently
found in developing countries, the price mechanism does not function well enough
to convey appropriate signals regarding supply and demand conditions. As such,
government intervention in the allocation of resources is unavoidable so that overall
production capacity can be expanded more rapidly than when the price mechanismalone
plays a major role in resource allocation. In addition, monetary expansion may be
considered a necessary by-product of the growth policy pursuing structural changes.
In any case, the “structuralists vs monetarists” debate continues. For instance, Faria and
Carneiro (2001) investigate the relationship between in?ation and output in the context
of an economy facing persistent high in?ation and they ?nd that in?ation does not affect
real output in the long run, but that in the short-run in?ation negatively affects output.
Finally, scholars such as Sidrauski (1967) suggest that there is no relationship between
in?ation and economic growth, supporting the hypothesis of superneutrality of money.
Empirical studies arrive at various conclusions, from papers ?nding no relationship
among the variables to studies asserting the existence of a nonlinear relationship.
De Gregorio (1993) studies 12 Latin-American countries in the period 1950-1985 and
?nds a signi?cant and negative relationship between economic growth and in?ation.
Barro (1995) studies panel data from 100 countries during the period 1960-1990 and
suggest a negative relationship. However, Bruno and Easterly (1998) do not agree with
these results, asserting that the negative relationship is due to a bias in the sample, they
argue that Barro included countries with in?ation larger than 40 percent and these
outliers potentially dominated the results. Fischer (1993) asserts that in?ation, the ?scal
de?cit and exchange rate distortions negatively impact on economic growth. Andres
and Hernando (1999) analyze the OECDcountries in1960-1992 andthey ?nd that current
in?ation has never been correlated positively with income per capita in the long run,
and most of the time the relationship is negative for long periods of time.
JFEP
1,3
248
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
On the other hand, Sarel (1995) asserts that there is a nonlinear relationship between
in?ation and economic growth. Using 87 countries, he ?nds the existence of an in?ation
threshold of 8 percent. Above the threshold there is a negative relationship between
in?ation and economic growth, whereas under the threshold there is a positive but not
signi?cant relationship. Judson andOrphanides (1996) divide Sarel’s sample of countries
into three groups, and they ?nd similar results to Sarel, ?nding a threshold of 10 percent.
Ghosh and Phillips (1998a, b) study 145 countries in the period 1960-1990 again ?nding
similar results. Paul et al. (1997) study 70 countries (of which 48 are developing
economies) for the period 1960-1989. They ?nd no causal relationship between in?ation
and economic growth in 40 percent of the countries, bidirectional causality among
20 percent of the countries, and unidirectional causality for the rest (either in?ation to
growth or vice versa). The relationship was found to be negative in some cases and
positive in others. Burdekin et al. (2004), studying a similar sample, highlight that
nonlinearities behave quite differently for industrial economies than they appear to
behave for developing countries. They ?nd that the threshold at which in?ation ?rst
begins to seriously and negatively affect growth is around 8 percent for industrial
economies but around 3 percent for developing countries. Loungani and Sheets (1997)
study 26 transition economies during the period 1991-1994. In?ation was high in the
period and they ?nd a negative relationship between in?ation and economic growth.
Mexico, similar to the rest of Latin America, has a history of long periods of high
in?ation. According to Gagnon (2009) Mexico suffered three digit in?ation rates in the
late 1980s. In 1994 in?ation was successfully stabilized below10 percent. However, after
the collapse of the exchange rate in 1994, in?ation peaked at 92 percent in April 1995 and
real output per capita contracted by 9.5 percent in 1995. Mexico would wait until 1998 for
real gross domestic product (GDP) per capita to surpass its 1994 level and until 1999 for
in?ation to settle below 10 percent.
The Mexican case has been studied by several scholars. Mendoza (1998) ?nds
that in?ation has had no effect on Mexico’s long-run economic growth. Katz (2002)
asserts that in?ation is dangerous for Mexico’s economic growth, identifying a negative
correlation between the variables in Mexico during the period 1970-2002. Acevedo (2006)
applies the nonlinear approach proposed by Sarel (1995) to the Mexican economy.
He ?nds similar results to Sarel (1995) and Judson and Orphanides (1996), showing
evidence for a threshold of 8.1 percent for Mexico, where in?ation rates higher than this
level severely damage Mexico’s economic performance.
Grier and Grier (2006) present evidence on the real effects of in?ation and in?ation
uncertainty on Mexican output growth. Their main ?ndings are as follows:
.
in?ation uncertainty has a negative and signi?cant effect on growth;
.
once the effect of in?ation uncertainty is accounted for, lagged in?ation does not
have a direct negative effect on output growth; and
.
however, as predicted by Ball (1992), higher average in?ation raises in?ation
uncertainty, and the overall net effect of average in?ation on output growth in
Mexico is negative.
That is, the average in?ation rate is harmful to Mexican growth due to its impact on
in?ation uncertainty.
It seems, therefore, that there is no theoretical consensus on exactly how in?ation
affects economic performance. Much of the empirical literature searches for a negative
In?ation and
Mexican
economic growth
249
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
in?uence of in?ation on growth, yet many economic theories predict neutrality or even a
positive effect of average in?ation on economic performance. So, the aimof this paper is
to study the nature of the relationship between in?ation and economic growth and
speci?cally, this paper focuses on two questions[1]:
Q1. Is there any long-run relationship between economic growth and in?ation in
Mexico?
Q2. Is there a statistically signi?cant threshold level of in?ation above which
in?ation affects growth differently than at lower in?ation rates in Mexico?
Motivated by these questions, the paper ?rst examines cointegration techniques and
proceeds, second, to threshold estimations (Hansen, 1999, 2000). To study the threshold
level of in?ation, we adopt the econometric technique developed by Khan and Senhadji
(2001) analyzing a nonlinear or structural break effect that examines whether the impact
of in?ation on economic growth could be positive up to a certain threshold level, but
beyond this level the effect turns negative. The model provides appropriate procedures
for estimation and inference. Using the nonlinear least squares (NLS) estimation
technique, Khan and Senhadji (2001) estimated that the threshold levels for industrial
countries and developing countries were at 1-3 and 11-12 percent, respectively[2].
We found a threshold level of 9 percent of in?ation above which the effect of in?ation is
negative and statistically signi?cant. But, in?ation has a negative effect on long-run
growth only if the level of in?ation is above such a threshold level. We found a
statistically signi?cant long-run negative relationship between in?ation and economic
growth. Our results therefore corroborate the results produced by De Gregorio (1993),
Barro (1995) and Katz (2002).
The remainder of the paper is organized as follows. In Section 2, we introduce the data
set and the applied method. Section 3 presents the estimated long-run relationship using
the cointegration analysis. In Section 4, we estimate the in?ation threshold by applying
the technique suggested by Khan and Senhadji (2001). Finally, in Section 5, we draw
some concluding remarks.
2. Data set and methodology
We consider annual data of real GDP (Y), consumer price index (CPI) and investment
share (I) in 2000 constant prices (Mexican local currency) for the period 1970-2007.
In?ation (p) was computed as the growth rate of the CPI; both data on the growth rate of
real GDP and CPI were obtained from World Economic Outlook (WEO) database. Data
on the investment share of real GDP was obtained fromthe Penn World Tables 6.2. Data
on population (annual growth rate) (POP) was obtained fromOECDstatistics portal. We
applied a log transformation, since log transformations help, at least partially, to
eliminate the strong asymmetry in the distribution of in?ation (Sarel (1995) and Ghosh
and Phillips (1998a, b)).
We consider two econometric models to obtain the empirical results:
(1) In the ?rst part of the paper, we examine the short- and long-run relationships
between real GDP and p. First, we examine the rank of integration for the series
log (real GDP) and (p) using the augmented Dickey-Fuller (ADF) Dickey and
Fuller (1981) unit root and the Kwiatkowski-Phillips-Schmidt-Shin tests (KPSS).
Second, an unrestricted vector autoregressive (VAR) is estimated, which
JFEP
1,3
250
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
is known to be sensitive to the number of time lags (Banerjee et al., 1993).
To determine the appropriate number of time lags for the VAR, we employed the
Akaike information criterion (AIC) and the Schwartz Bayesian criterion. Third,
we examine the existence of a long-run relationship between GDP and p using a
vector error-correction model (VECM) after applying Johansen’s (1988, 1990, and
1995) cointegration technique. We conduct a test for weak exogeneity in order to
do inference. Finally, the causality between the variables is studied based on the
more robust Toda and Yamamoto (1995) Granger no-causality test which allows
the use of the Granger test in an integrated system.
(2) The second part of the paper attempts to ?nd the mechanism through which
in?ation nonlinearly affects long-run economic growth. Hence, we consider Khan
and Senhadji’s (2001) model to estimate a threshold level of in?ation above which
in?ation affects economic growth negatively.
3. Cointegration analysis
A spurious regression results when trending or unit root time series produce
non-stationary residuals, signi?cant ordinary least-squares (OLS) parameter estimates,
and a high-R
2
. Non-stationary residuals violate the standard assumptions that are
required to apply OLS methods. In this case, Phillips (1986) pointed out that
cointegration techniques must be applied. Recently, cointegration tests are widely used
in econometric practice. Cointegration tests suggest that restrictions be imposed on
multivariate or VAR models and can be used to test economic theories.
A?rst step in cointegration analysis is to study the stationarity of the series by using
unit root tests. Two popular techniques have been used: the ADFtest (Dickey and Fuller,
1981) and the KPSS test (Kwiatkowski, et al. 1992). These tests have been performed in
levels (i.e. log of real GDP and in?ation rate) as well as in ?rst differences. If the two time
series are integrated of the same order, then we can estimate a cointegrating regression.
Thus, we search for a long-run relationship between the two variables, followed by
applying a VECM to model the short-run dynamics. The model is represented in a
?rst-differenced error-correction form:
DY
t
¼ m þPY
t21
þ
X
i¼k21
i¼1
G
i
DY
t2i
þ1
i
z ð1Þ
where, Y ¼ (real GDP, in?ation) is a vector containing the variables and mis a vector of
constant terms. The matrix P conveys information about the long-run relationship
between the Y variables. The rank of P is the number of linearly independent and
stationary linear combinations of the variables.
McCallum (1984) asserts that incorrect signs can be produced if exogeneity is not
studied. To apply inference techniques, we must test weak exogeneity. Finally,
a modi?ed version of the Granger causality test is applied in order to analyze causality
between the variables.
3.1 Empirical results
A ?rst step in cointegration analysis is to study the stationarity of the series by using
unit root tests, such as the ADF and the KPSS tests. The null hypothesis of the KPSS test
is stationarity, complementing the ADF test. However, the ADF test has low power
against stationary near unit root processes. Nevertheless, Tables I and II show unit root
In?ation and
Mexican
economic growth
251
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
V
a
r
i
a
b
l
e
y
p
D
y
D
p
U
n
i
t
r
o
o
t
t
e
s
t
A
D
F
K
P
S
S
A
D
F
K
P
S
S
A
D
F
K
P
S
S
A
D
F
K
P
S
S
T
r
e
n
d
a
n
d
c
o
n
s
t
a
n
t
2
2
.
9
7
(
2
3
.
5
4
)
0
.
1
4
(
0
.
1
5
)
2
2
.
3
2
(
2
3
.
5
4
)
0
.
1
6
(
0
.
1
5
*
)
2
4
.
4
6
(
2
3
.
5
4
)
*
0
.
1
1
(
0
.
1
5
)
2
5
.
8
4
(
2
3
.
5
4
)
*
0
.
5
0
(
0
.
1
5
)
*
C
o
n
s
t
a
n
t
2
2
.
2
0
(
2
2
.
9
5
)
0
.
7
3
(
0
.
4
6
)
*
2
2
.
0
6
(
2
2
.
9
5
)
0
.
2
3
(
0
.
4
6
)
2
4
.
2
3
(
2
2
.
9
5
)
*
0
.
2
8
(
0
.
4
6
)
2
5
.
7
6
(
2
2
.
9
5
)
*
0
.
3
5
(
0
.
4
6
)
W
i
t
h
o
u
t
t
r
e
n
d
,
c
o
n
s
t
a
n
t
2
5
.
9
3
(
2
1
.
9
5
)
2
1
.
3
7
(
2
1
.
9
5
)
2
2
.
6
8
(
2
1
.
9
5
)
*
2
5
.
8
5
(
2
1
.
9
5
)
*
N
o
t
e
s
:
*
N
u
l
l
h
y
p
o
t
h
e
s
i
s
r
e
j
e
c
t
i
o
n
a
t
5
p
e
r
c
e
n
t
;
c
r
i
t
i
c
a
l
v
a
l
u
e
s
i
n
p
a
r
e
n
t
h
e
s
i
s
Table I.
ADF and KPSS unit root
tests
JFEP
1,3
252
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
tests for the variables in levels and in differences, the variables are expressed in
logarithmic form.
According to the tests, the time series are integrated processes of ?rst order, that is
they are I(1). Hence, we have to study the existence of a cointegrating relationship.
Banerjee et al. (1993) highlight that searching for a cointegration relation is equivalent to
searching for a statistical equilibriumbetween variables tending to growover time. The
discrepancy of this equilibrium can be modeled by a VECM (equation (1)). The VEC
model shows howthe variables return to the equilibriumafter suffering a shock. In order
to obtain the optimal VEC model, we applied the minimum AIC criterion. We estimate a
VAR with one lag (according to the minimum AIC) and test the cointegration
relationship. Table II indicates that one co-integrating relationship is found.
To do inference, we need to test the weak exogeneity of real GDP. The x
2
-statistic is
1.01 producing a p-value of 0.31. Therefore, we cannot reject the hypothesis that real
GDP is weakly exogenous at a 5 percent level of signi?cance.
In the second stage, the VEC model is employed to assess whether the economy
approaches equilibrium in the long run and to analyze the short-run dynamics of the
cointegrated time series variables. The VECis internally consistent if the two time series
variables are cointegrated of the same order or if they are stationary. Equation (2) shows
the long-run estimated relationship:
y ¼ þ7:02 21:56p ½231:53? ½3:01? ð2Þ
Equation (2) shows that the elasticity for in?ation-GDP was approximately 21.56
between 1970 and 2007. This elasticity means that an increase of the annual in?ation of
1 percent produced a decrease in GDP of more than 1.5 percent. This result coincides
with De Gregorio (1993), Barro (1995) and Katz (2002). We mentioned some work
suggesting that the negative relationship is found when in?ation is high. During most of
the period, Mexico suffered high in?ation and the cointegration relationship could
suggest a negative relationship. Since in?ation is weakly exogenous, we can perform
inference. We can say, therefore that high in?ation had a negative impact on economic
growth. Notice though that our result differs from the case shown by Mallik and
Chowdhury (2001). They examine the relationship between in?ation and GDP growth
for four South Asian countries (Bangladesh, India, Pakistan, and Sri Lanka). They ?nd
evidence of a long-run positive relationship between GDP growth rate and in?ation for
all four countries. Predominantly, these countries have not had high-in?ation crises
(except Bangladesh during 1972-1974 only); their in?ation rates of 7-10 percent can thus
be regarded as moderate.
Hypothesis Trace statistic CV at 0.05 p-value
None
a
21.44 20.26 0.034
At most 1 4.73 9.16 0.314
Hypothesis Max. Eig. Stat. CV at 0.05 p-value
None
a
16.71 15.89 0.037
At most 1 4.73 9.16 0.314
Note:
a
Indicates rejection of the null hypothesis at 5 percent
Source: Own calculations
Table II.
Johansen cointegration
test
In?ation and
Mexican
economic growth
253
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Though cointegration is required to examine the relationships, it does not indicate
the direction of the causal relationship. Granger (1988) proposed a test to study
causality, subsequently called Granger causality. But Granger causality does not imply
causality in a philosophical sense; rather it should be understood instead as a kind
of predetermination among variables. Dynamic Granger causality can be captured in the
VAR model. Notice that, the variables are integrated of the same order, and applying
the standard Granger causality test would be not appropriated. In this vein, Toda and
Yamamoto (1995) suggest an alternative procedure. When the variables are integrated,
they propose estimating a VAR model with (k þ d
max
) lags, where k is the standard
optimal number of lags and d
max
is the maximal order of integration that we suspect
might occur in the process. Once the VAR is estimated, we may test Granger causality
using only the ?rst k lags. For instance, if we consider the following equation from a
VAR model:
p
t
¼ g
0
þg
1
Y
t21
þg
2
Y
t22
þg
3
p
t21
þg
4
p
t22
ð3Þ
where, k ¼ 1 was selected according the minimum AIC and d
max
¼ 1, the null
hypothesis of non-causality from Y to p should be:
H
0
: g
1
¼ 0 ð4Þ
Equation (4) means:
H
0
real GDP does not Granger-cause in?ation
The hypothesis is tested using the Wald test. Toda and Yamamoto (1995) assert that
the Wald and LR tests are asymptotically equivalent in the present situation. Table III
shows the results for all the variables.
According to the Wald statistic, we cannot reject the null hypothesis of no causality.
Therefore, we do not ?nd evidence of Granger-causality among the variables.
These ?ndings have important policy implications. Even if we do not ?nd any
direction in Granger causality, the weak exogenous test indicates that high levels of
in?ation were harmful to economic growth in the period considered. Hence, our evidence
corresponds to the monetarist position, and caution is needed since periods of higher
in?ation may produce negative in?ation elasticities.
4. Threshold analysis
Khan and Senhadji (2001) calculate growth rates of macroeconomic variables using a log
transformation because a log transformation provides a best ?t in the class of nonlinear
models. Although the growth rates of all these variables are calculated using log
transformation method, a lot of volatility remains in the data (Figure 1).
Null hypothesis Wald-statistic p-value
Y does not Granger cause P 0.05 0.83
P does not Granger cause Y 0.08 0.78
Notes: We applied a VAR with k þ dmax ¼ 1 þ 1; p-values correspond to the x
2
distribution with
two degree of freedom
Source: Toda and Yamamoto (1995)
Table III.
Granger causality test
JFEP
1,3
254
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Therefore, in order to smooth out business cycle ?uctuations, the data set is smoothed
using the Hodrick-Prescott ?lter (Figure 2).
To test for the existence of a threshold effect, the following model was estimated:
y
t
¼ b
0
þb
1
logðp
t
Þ þb
2
d
p
*
t
½logðp
t
Þ 2logðp
*
Þ? þu
0
X
t
þ e
t
ð5Þ
d
p
*
t
¼
1 if p
t
. p
*
0 if p
t
# p
*
8
<
:
t ¼ 1; . . . ; T
where, y
t
is the growth rate of real GDP, p
t
is the in?ation rate based on the CPI, p
*
is the
thresholdlevel of in?ation, d
p
*
t
is a dummyvariable that takes a value of one for in?ation
levels greater than p
*
percent and zero otherwise, X
t
is a vector of control variables
which includes investment as a share of GDP and the population growth rate. Our study
Figure 1.
Log transformation
of series
12
8
4
0
–4
–8
1970 1975 1980 1985 1990 1995 2000 2005
LNINF GDP
Figure 2.
Series smoothed
50
40
30
20
10
0
1970 1975 1980 1985 1990 1995 2000 2005
GDP filter
Invest filter
Infilter
Pop filter
In?ation and
Mexican
economic growth
255
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
is based on a four-variable model consisting of economic growth, in?ation, population,
and investment[3]. The parameter p
*
represents the threshold in?ation level with the
property that the relationship between output growth and in?ation is given by:
.
low-in?ation: b
1
; and
.
high in?ation: b
1
þ b
2
.
High in?ation means that when the long-run in?ation estimate is signi?cant then b
1
and
b
2
would be summed to see their impact ongrowth and that would produce the threshold
level of in?ation. By estimating regressions for different values of p
*
, chosen in
ascending order (i.e. 1, 2 and so on), the optimal value of p
*
is obtained by ?nding the
value that minimizes the residual sumof squares (RSS)[4]. In?ation beyond this level has
a signi?cant and negative impact on economic growth. This procedure has become
widely accepted in the literature. However, this process is tedious since it requires
estimating the equation several times for different values of p
*
. Nevertheless, statistical
rigor requires that it be done.
4.1 Empirical results
By using NLS to estimate equation (5), we can obtain the exact value of the threshold
in?ation level. The minimum RSS of each regression considers p
*
¼ 1 to p
*
¼ 13.
However, the values of p
*
when RSS is minimized ranges from 7-11 percent. The
estimated results are reported in Table IV.
According to the p-values, the estimator parameters are statistically signi?cant at very
lowvalues. Therefore, in the estimation process the threshold level of in?ation is observed
at the 9 percent level where RSS is minimized. Hence, while in?ation belowthis threshold
level has no signi?cant effect on economic growth (i.e. statistically insigni?cant at
5 percent level), in?ation rates above the threshold have a signi?cant and negative effect
on economic growth.
Fromthe estimated results, at lower threshold in?ation levels (p
*
less than 9 percent)
there is no statistically signi?cant relationship between the dummy of threshold level of
in?ation and economic growth. As p
*
is greater than9 percent, a statisticallysigni?cant
relationship is observed between economic growth and the dummy of threshold level
of in?ation which continues upto an11-percent in?ation rate, whichmeans that in?ation
has a negative statistically signi?cant effect on Mexican economic growth for in?ation
levels greater than 9 percent.
Since 1999, the Banco de Me´xico announced a mediumtermin?ation objective, and since
2000 the Central Bank publishes quarterly in?ation reports to monitor the in?ationary
process, to analyze in?ationprospects andto discuss the conduct of monetarypolicyandthe
balance of risks for future in?ation (BANXICO, 2008). Currently, Mexico is considered to
have in place the main components of an in?ation targeting framework: an independent
monetary authority (since 1993) that has in?ation as its only policy objective, a ?exible
exchange rate regime, the absence of other nominal anchors and a “transparent” framework
for the implementation of monetary policy (Schmidt-Hebbel and Werner, 2002).
In this vein, our results follow from the in?ation targeting aims that The Mexican
Central Bank de?ned in a series of annual in?ation targets since 1999. The target ceiling
started at 13 percent for 1999, decreased to 10 percent by 2000, and was followed
by additional decreases to 9 percent for the 2001 and 2002, 4.05 percent in 2006, and
3.76 percent in 2007.
JFEP
1,3
256
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
4.1.1 2SLS analysis. In equation (5), in?ation may not be exogenous and consequently
the coef?cient estimates may be biased. As Fischer (1993) pointed out, the causality is
more likely to be unidirectional from in?ation to growth, in which case the problem of
simultaneity bias may not be very important. Consequently, a Granger Causality test is
applied to measure the linear causation between in?ation and economic growth. Table V
shows that the null hypothesis is not rejected, which means that GDP (economic growth)
Granger causes in?ation. The causality between two variables is unidirectional since the
null hypothesis that in?ation causes output growth is not rejected at a 5 percent level of
signi?cance, which supports the hypothesis that there is no feedback from output
growth to in?ation.
Hence, in?ation causes economic growth (in the sense of either a positive sign or
negative sign on its coef?cient) but economic growth is not only explained by in?ation
rate, then the model could be incorrectly speci?ed, implying a speci?cation bias in the
Null hypothesis F-statistic p-value
In?ation does not Granger cause economic growth 1.50237 0.23837
Economic growth does not Granger cause in?ation 6.44918 0.00455
*
Note:
*
Indicates rejection of the null hypothesis at 5 percent
Table V.
Pairwise Granger
causality test
p
*
Variable Coef?cient SE t-statistic Prob. RSS
C 226.57851 2.192942 212.12002 0.0000
Log(p) 20.488365 0.073403 26.653202 0.0000
7 percent Break7 20.092338 0.031913 22.893442 0.0067 10.40
Pop 1.503096 0.064185 23.41810 0.0000
Inv 9.853121 0.742647 13.26758 0.0000
C 226.40593 2.110052 212.51435 0.0000
Log(p) 20.545863 0.079201 26.892144 0.0000
8 percent Break8 20.090380 0.026949 23.353742 0.0020 1.78
Pop 1.521239 0.062812 24.21894 0.0000
Inv 9.822111 0.711538 13.80405 0.0000
C 226.74910 1.963577 213.62264 0.0000
Log(p) 20.569815 0.075737 27.523557 0.0000
9 percent Break9 20.086855 0.021997 23.948469 0.0004 1.62
Pop 1.494305 0.058858 25.38810 0.0000
Inv 9.970370 0.660360 15.09839 0.0000
C 226.81594 1.969623 213.61476 0.0000
Log(p) 20.574085 0.077255 27.431035 0.0000
10 percent Break10 20.075255 0.019342 23.890821 0.0005 1.64
Pop 1.493487 0.059123 25.26058 0.0000
Inv 9.996084 0.662225 15.09470 0.0000
C 227.70299 2.028275 213.65840 0.0000
Log(p) 20.553976 0.082645 26.703076 0.0000
11 percent Break11 20.058091 0.017901 23.245056 0.0027 1.81
Pop 1.496378 0.062279 24.02695 0.0000
Inv 10.30664 0.682291 15.10593 0.0000
Note: Where break number is de?ned by ½logðp
t
Þ 2logðp
*
Þ?
Table IV.
Test results for the
threshold effects
estimation
In?ation and
Mexican
economic growth
257
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
estimated model (5), that is, exclusion of other relevant variables for a growth equation.
To check this bias, in?ation, real GDP growth, investment, and population growth are
used as instruments[5]. In order to control for endogeneity, at least partially, the model is
now estimated using two-stage least squares (2SLS), where the set of instruments
includes the lag of in?ation, the lag of real GDP, the lag of investment and the lag of the
population rate[6]. The results of 2SLS also suggest a 9 percent threshold in?ation level
(Table VI).
Both estimated models (Table IV (NLS) and Table VI (2SLS)) provide identical
threshold levels of in?ation and the values of estimated coef?cients also remain close
in both models. Both results indicate a 9 percent threshold in?ation level for economic
growth in Mexico. This threshold level lies within the range found by Khan and Senhadji
(2001), and it can be interpreted as the threshold level for a transition economy. These
results are similar to Sarel (1995), Judson andOrphanides (1996) in the international case,
and Acevedo (2006) who ?nds a threshold of 8.1 percent for Mexico, where in?ation rates
greater than this level would severely damage Mexico’s economic performance.
5. Concluding remarks
The main objective of this paper was to examine whether a relationship exists between
economic growth and in?ation and, if so, its nature. Even if there is neither a theoretical
p
*
Variable Coef?cient SE t-statistic Prob. RSS
C 227.60274 1.065922 22.589566 0.09930556
Log(p) 20.719575 0.358756 22.005752 0.37083333
7 percent Break7 20.677079 0.460916 21.468986 1.05277778 25.14
Pop 1.223486 0.452033 2.706632 0.075
Inv 9.976084 3.516970 2.836557 0.0078
C 220.07569 4.742451 24.233189 0.0002
Log(p) 21.041340 0.211722 24.918432 0.0000
8 percent Break8 20.317888 0.099192 23.204783 0.0031 5.18
Pop 1.681314 0.151782 1.107718 0.0000
Inv 7.704655 1.588038 4.851682 0.0000
C 220.76328 3.717135 25.585830 0.0000
Log(p) 21.020460 0.163859 26.227659 0.0000
9 percent Break9 20.225560 0.055577 24.058510 0.0003 3.29
Pop 1.630786 0.120352 13.55017 0.0000
Inv 8.031724 1.241283 6.470503 0.0000
C 220.63593 3.729246 25.533540 0.0000
Log(p) 21.045673 0.168227 26.215838 0.0000
10 percent Break10 20.196727 0.048571 24.050300 0.0003 3.31
Pop 1.636986 0.120465 13.58889 0.0000
Inv 7.999773 1.244817 6.426465 0.0000
C 218.82909 3.941537 24.777092 0.0000
Log(p) 21.194979 0.200614 25.956612 0.0000
11 percent Break11 20.166932 0.042509 23.927015 0.0004 3.52
Pop 1.770391 0.126191 14.02943 0.0000
Inv 7.461593 1.309084 5.699856 0.0000
Note: Where break number is de?ned by ½logðp
t
Þ 2logðp
*
Þ?
Table VI.
Test results for
the threshold effects
of 2SLS estimation
JFEP
1,3
258
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
nor an empirical consensus about the relationships between these variables, some studies
seem to suggest that an in?ation rate threshold exists above which the relationship
between in?ation and economic growth is negative and below which the relationship
is positive or zero. This would mean that high levels of in?ation may harm the economy.
However, setting zero in?ation could also be harmful when production is growing.
According to Akerlof et al. (1996) ?rms cannot support shocks of demand under zero
in?ation. These shocks could then produce inef?cient, low levels of employment. This
suggests that there may be a minimum ef?cient level of in?ation. Furthermore, Sinn and
Reuter (2001) assert that estimating a minimum level of in?ation is an important step in
in?ation targeting to prevent de?ation.
We used the cointegration technique, and found a signi?cant and negative long-run
relationship between in?ation and economic growth for the Mexican economy.
In addition, we found that in?ation is weakly exogenous. In the period 1970-2007 real
GDP was elastic with respect to in?ation, and therefore, considering the estimated
coef?cient, an increase of 1 percent on in?ation produces a decrease of 1.5 percent on
real GDP. Since, for most of the period under consideration Mexico experienced
in?ation rates higher than 10 percent, this result is consistent with most of the research
suggesting that high levels of in?ation produce a negative effect on economic growth.
Using a more robust Granger causality test, we found no predetermination between the
variables. However, the weak exogeneity of in?ation allows us to do inference and
shows that during the period 1970-2007 increasing (decreasing) in?ation was harmful
(bene?cial) to the economy.
As mentioned before, Mexico has an in?ation targeting policy. Thus, our second
analysis could be useful for policymakers in providing some clue in setting an optimal
in?ation target. For instance, the Mexican Central Bank could apply an expansionary
monetary policy for supporting economic growth until the in?ation rate does not
exceed the threshold level. In fact, our threshold analysis suggests that if the in?ation
rate exceeds 9 percent, then Mexico’s current favorable economic performance might
be jeopardized. Mishkin (2000) asserts that one of the advantages of the in?ation
targets is to respond to domestic shocks. International lending agencies gave policy
advice to Mexico, outlining how it should attempt to reduce in?ation to a very low level
to get positive effects on economic growth. The relevant question becomes what should
this level of low-in?ation be?
To conclude, as is now well-known, the Mexican economy has been devastated by
the current economic crisis. Mexico has experienced in?ation in excess of 5 percent
because of increasing costs for food and for fuel, with commensurate decreases in
demand for basic commodities affecting GDP. But, though the in?ation target is now
3 percent, Mexico is experiencing in?ation around 5 percent (4.89 percent at most recent
estimates). As a consequence, of the crisis, Mexico is unlikely to attain in?ation of
3 percent until later in 2010, when optimistic forecasts show that its growth may also
return to 3 percent or higher. Though debates about in?ation targeting, Mexico will
continue to feel the brunt of the economic crisis in conjunction with the other problems
endemic to its economy. That is, according to our results and the empirical evidence in
Mexico, the in?ation targeting is far from being a serious problem to attain a sustained
economic growth in the country. Nowadays, other problems like corruption, drug
traf?cking, and unemployment are crucial barriers for the Mexican growing.
In?ation and
Mexican
economic growth
259
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Notes
1. We chose this focus since the Banco de Mexico (or Mexican Central Bank) is required by the
Constitution (Article 28, paragraph 6) to achieve price stability as its primary objective.
According to the bank’s 1994 annual report, in?ation abatement had been given special
emphasis in Mexican economic policy since the end of 1987. Reducing in?ation was viewed as
a necessary condition for attaining social justice and long-run economic growth (Banco de
Me´xico 1995, p. 50).
2. Khan and Senhadji (2001) used an unbalanced panel with 140 countries over 40 years to
precisely examine the nonlinear relationship between in?ation and growth. The estimate of
the threshold level was 1-3 percent for industrial countries and 11-12 percent for developing
countries.
3. Population growth rate and investment are used as control variables. The reason for choosing
these variables is their authenticity in empirical literature on growth. Solow (1956) and Swan
(1956) who developed the ?rst neo-classical models of growth, take the rate of growth of
population as one of exogenous variables in their model to show that the faster the rate of
population growth, the poorer the country. Fischer (1993) includes investment in his model to
show that in?ation reduces growth by reducing investment and productivity growth.
Moreover, Mankiwet al. (1992) also include investment growth and population growth in their
growth model.
4. As KhanandSenhadji (2001) pointedout testing for no thresholdeffects means H
0
: b
2
¼ 0, but
under the null hypothesis, the threshold p
*
is not identi?ed, so classical tests, such as the
t-test, have nonstandard distributions. They suggest following the Hansen (1996, 1999)
methodology to simulate an asymptotic distribution of the likelihood ratio test of H
0
, which
implies to minimize the RSS.
5. For instance, investment (as share of GDP) is also likely to be endogenous to growth.
6. These are valid instruments only if the error term in equation (5) is not autocorrelated.
The DW and Breusch-Godfrey tests reject autocorrelation at 5 percent for all estimations.
References
Acevedo, E. (2006), “In?acio´n y crecimiento econo´mico en Me´xico: una relacio´n no lineal”,
Econom? ´a Mexicana Nueva E
´
poca, Vol. 15 No. 2, pp. 199-249.
Akerlof, G., Dickens, W. and Parry, G. (1996), “The macroeconomics of low in?ation”, Brookings
Papers on Economic Activity, Vol. 27 No. 1, pp. 1-76.
Andres, J. and Hernando, I. (1999), “Does in?ation harm economic growth? Evidence for the
OECD countries”, in Feldstein, M. (Ed.), The Costs and Effects of Price Stability,
The University of Chicago Press, Chicago, IL, pp. 315-41.
Ball, L. (1992), “Why does high in?ation raise in?ation uncertainty?”, Journal of Monetary
Economics, Vol. 29 No. 3, pp. 371-88.
Banco de Me´xico (1995), Informe Annual: 1994, Banco de Me´xico, Me´xico City.
Banerjee, A., Dolado, J., Galbraith, J. and Hendry, D. (1993), Co-integration, Error Correction, and
the Econometric Analysis of the Non-Stationary Data, Oxford University Press, Oxford.
BANXICO (2008), Quarterly In?ation Reports, The Central Bank of Mexico, BANXICO, Mexico
City, available: www.banxico.org.mx/publicaciones/JSP/b_in?ationreport.jsp
Barro, R.J (1995), “In?ation and economic growth”, NBER Working Papers No. 5326,
National Bureau of Economic Research, Cambridge, MA.
Bruno, M. and Easterly, W. (1998), “In?ation crises and long-run growth”, Journal of Monetary
Economics, Vol. 41 No. 1, pp. 3-26.
JFEP
1,3
260
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Burdekin, R., Denzau, A., Keil, M., Sitthiyot, T. and Willett, T. (2004), “When does in?ation hurt
economic growth? Different nonlinearities for different economies”, Journal of
Macroeconomics, Vol. 26 No. 3, pp. 519-32.
Cooley, T. and Hansen, G. (1991), “The in?ation tax in a real business cycle model”, American
Economic Review, Vol. 79 No. 4, pp. 733-48.
De Gregorio, J. (1993), “In?ation, taxation and long-run growth”, Journal of Monetary Economics,
Vol. 31 No. 3, pp. 271-98.
De Gregorio, J. (1996), “In?ation, growth, and central banks: theory and evidence”, Policy
Research Working Paper Series No. 1575, The World Bank, Washington, DC.
Dickey, D.A. and Fuller, W.A. (1981), “Likelihood ratio statistics for autoregressive time series
with a unit root”, Econometrica, Vol. 49 No. 4, pp. 1057-72.
Faria, R.J. and Carneiro, F. (2001), “Does high in?ation affect growth in the long and short run?”,
Journal of Applied Economics, Vol. 4 No. 1, pp. 89-105.
Fischer, I. (1926), “A statistical relationship between unemployment and price changes”,
International Labor Review, Vol. 13.
Fischer, S. (1993), “The role of macroeconomic factors in economic growth”, Journal of Monetary
Economics, Vol. 32 No. 2, pp. 485-512.
Fischer, S. and Modigliani, F. (1980), “Towards and understanding of the real effects and costs of
in?ation”, NBER Working Papers No. 0303, National Bureau of Economic Research,
Cambridge, MA.
Gagnon, E. (2009), “Price setting during low and high in?ation: evidence from Mexico”,
The Quarterly Journal of Economics, Vol. 124 No. 3, pp. 1221-63.
Ghosh, A. and Phillips, S. (1998a), “In?ation, disin?ation, and growth”, IMF Working Paper
No.WP/98/68, IMF, Washington, DC.
Ghosh, A. and Phillips, S. (1998b), “Warning: in?ation may be harmful to your growth”, IMF
Staff Papers, Vol. 45, IMF, Washington, DC.
Granger, C. (1988), “Some recent developments in a concept of causality”, Journal of
Econometrics, Vol. 39 Nos 1/2, pp. 199-211.
Grier, K. and Grier, R. (2006), “On the real effects of in?ation and in?ation uncertainty in Mexico”,
Journal of Development Economics, Vol. 80 No. 2, pp. 478-500.
Hansen, B. (1996), “Inference when a nuisance parameter is not identi?ed under the null
hypothesis”, Econometrica, Vol. 64 No. 2, pp. 413-30.
Hansen, B. (1999), “Threshold effects in non-dynamic panels: estimation, testing, and inference”,
Journal of Econometrics, Vol. 93 No. 2, pp. 345-68.
Hansen, B. (2000), “Sample splitting and threshold estimation”, Econometrica, Vol. 68
No. 3, pp. 575-604.
Johansen, S. (1988), “Statistical analysis of cointegration vectors”, Journal of Economic Dynamics
and Control, Vol. 12 Nos 2/3, pp. 231-54.
Johansen, S. (1995), Likelihood-based Inference in Cointegrated Vector Autorregressive Models,
Oxford University Press, Oxford.
Johansen, S. and Juselius, K. (1990), “Maximum likelihood estimation and inference on
cointegration with applications to the demand for money”, Oxford Bulletin of Economics
and Statistics, Vol. 52 No. 2, pp. 169-210.
Judson, R. and Orphanides, A. (1996), “In?ation, volatility and growth, ?nance and economics”,
Discussion Paper No. 96-19, Board of Governors, Federal Reserve of United States,
Washington, DC.
In?ation and
Mexican
economic growth
261
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Katz, I. (2002), “In?acio´n, crecimiento, pobreza y desigualdad en Me´xico”, en La in?acio´n
en Me´xico, Gaceta de Econom? ´a, Instituto Tecnolo´gico Auto´nomo de Me´xico, Mexico City.
Khan, M.S. and Senhadji, A.S. (2001), “Threshold effects in the relationship between in?ation and
growth”, IMF Staff Papers, Vol. 48 No. 1.
Kwiatkowski, D., Phillips, P., Schmidt, P. and Shin, Y. (1992), “Testing the null hypothesis
of stationarity against the alternative of a unit root”, Journal of Econometrics, Vol. 54
Nos 1-3, pp. 159-78.
Loungani, P. and Sheets, N. (1997), “Central bank independence, in?ation, and growth in
transition economies”, Journal of Money, Credit & Banking, Vol. 29 No. 3, pp. 381-99.
McCallum, B. (1984), “On low-frequency estimates of long-run relationships in macroeconomics”,
Journal of Monetary Economics, Vol. 14 No. 1, pp. 3-14.
Mallik, G. and Chowdhury, A. (2001), “In?ation and economic growth: evidence from South
Asian countries”, Asian Paci?c Development Journal, Vol. 8 No. 1, pp. 123-35.
Mankiw, N., Romer, D. and Weil, D. (1992), “A contribution to the empirics of economic growth”,
The Quarterly Journal of Economics, Vol. 107 No. 2, pp. 407-37.
Mendoza, M. (1998), “In?acio´n y crecimiento econo´mico en Me´xico, Monetaria”, revista
del CEMLA, Vol. 21 No. 2.
Mishkin, F. (2000), “In?ation targeting in emerging market countries”, The American Economic
Review, Vol. 90 No. 2, pp. 105-9.
Mundell, R. (1965), “Growth, stability and in?ationary ?nance”, Journal of Political Economy,
Vol. 73 No. 2, pp. 97-109.
Okun, A. (1962), “Potential GNP: its measurement and signi?cance”, in Pechman, J. (Ed.),
Economics for Policymaking, MIT Press, Cambridge, MA, 1983.
Paul, S., Kearney, C. and Chowdhury, K. (1997), “In?ation and economic growth: a multi-country
empirical analysis”, Applied Economics, Vol. 29 No. 10, pp. 1387-401.
Phillips, A. (1958), “The relationship between unemployment and the rate of change of money
wages in the United Kingdom, 1861-1957”, Economica, Vol. 25 No. 100, pp. 283-99.
Phillips, P. (1986), “Understanding spurious regressions in econometrics”, Journal of
Econometrics, Vol. 33 No. 3, pp. 311-40.
Sarel, M. (1995), “Nonlinear effects of in?ation on economic growth”, IMF Working Paper
Nos 95/56, International Monetary Fund, Washington, DC.
Schmidt-Hebbel, K. and Werner, A. (2002), “In?ation targeting in Brazil Chile and Mexico:
performance, credibility and exchange rate”, Central Bank of Chile Working Paper No. 171,
Central Bank of Chile, Santiago.
Sidrauski, M. (1967), “In?ation and economic growth”, Journal of Political Economy, Vol. 75 No. 6,
pp. 796-810.
Sinn, H. and Reuter, M. (2001), “The minimum in?ation rate for Euroland”, NBER Working
Paper, 8085, National Bureau of Economic Research, Cambridge, MA.
Solow, R. (1956), “A contribution to the theory of economic growth”, The Quarterly Journal of
Economics, Vol. 70 No. 1, pp. 65-94.
Swan, T. (1956), “Economic growth and capital accumulation”, Economic Record, Vol. 32
No. 63, pp. 344-61.
Tobin, J. (1965), “Money and economic growth”, Econometrica, Vol. 33 No. 4, pp. 671-84.
Toda, H. and Yamamoto, T. (1995), “Statistical inference in vector autorregressions with possibly
integrated processes”, Journal of Econometrics, Vol. 66 Nos 1/2, pp. 225-50.
JFEP
1,3
262
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Further reading
Engle, R. and Granger, C. (1987), “Co-integration and error correction: representation, estimation
and testing”, Econometrica, Vol. 55 No. 2, pp. 251-76.
Sweidan, O.D. (2004), “Does in?ation harm economic growth in Jordan? An econometric analysis
for the period 1970-2000”, International Journal of Applied Econometrics and Quantitative
Studies, Vol. 1 No. 2, pp. 41-66.
Corresponding author
Edgar J. Sa´nchez Carrera can be contacted at: [email protected]
In?ation and
Mexican
economic growth
263
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
6
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
doc_392831487.pdf