J. Risk Financial Manag. [613134]

J. Risk Financial Manag.
20
20
,
1
3
,
x; doi: FOR PEER REVIEW

www.mdpi.com/journal/
jrfm

The Impact

of USMCA on Stock Markets of the US, Canada, and
1

Mexico

2

3

Abstract

4

The study examines the impact
s

of the election of Donald J. Trump as the 45
th

President
5

of the United States and the subsequent events that
result in

the replacement of NAFTA
6

with the
United States
–
Mexico
–
Canada Agreement

(USMCA)
on the stock markets of the
7

U
S
, Canada
,

and

Mexico. The study looks at the
major

stock market indices
of the three
8

countries
.
The results indicate that

t
he
returns of stock indices on
Election Day

are
9

positive in the

US

and Canada
; however, more negative
accrued

for the subsequent
10

events,
in both of the

stock market
s
. In
the
case of Mexico,
most of the

sample
ind
ex
11

returns

are
negative
on the US presidential Election Day
followed by positive returns
for
12

all subsequent events
.
Our results are robust based on OLS and

counterchecked with

13

GARCH models.

14

15

Keywords
:

NAFTA, USMCA, presidential election

16

17

1.

Introduction

18

On November 9, 2016, Donald Trump got elected as the president of the United
19

States in an election that many observers felt was too close to call. Mr. Trump mentioned
20

his oppo
sition to the North American Free Trade Agreement (NAFTA) several times
21

during his election campaign. This agreement, which came into force on January 1,
22

1994, was a major
concern

during the election. President Trump announced on January
23

23,
2017
,

that t
he US would be withdrawing from NAFTA and mentioning that he would
24

negotiate a new trade agreement between the
US
, Canada
,

and

Mexico. After
several
25

months of negotiations, on September 30,
2018
,

the three countries finalized a new
26

agreement, the United States
–
Mexico
–
Canada Agreement (USMCA) to replace NAFTA.

27

The purpose of this paper is to
analyze

the impact
s

of this chain of events starting
28

with Mr. Trump’s surprise victory on the stock market returns of the U
S
, Canada
,

and

29

Mexico.
W
e
analyze

most of

the leading
stock
indices
in the

three. Our results
are robust
30

and
indicate that the major stock indices of
the U
S

and Canada
generally
have
positive
31

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

2

of
20

returns
on
Election Day

with later events having

predominantly

negative returns.
The

32

stock market

in Mexico, however,

experiences
negative returns on
the US

Election Day

33

and the returns turn into
positive

ones

when

the

renegotiation of
NAFTA

is announced
.
34

Positive

returns
accrue for

most

of the

indices
until the announcement of USMCA

35

creation in

Mexico
,
including

o
n the announcement
day
of USMCA

formation
.

36

When NAFTA came into effect in November 1993, several
studies examine the
37

impact
s

on the stock markets of the participating countries. Hanson and Song
(
1998)
38

report that
NAFTA does not favor
US

firms

but

favors
Mexican
ones
. The
y also find

that
39

the stock
returns
are

significantly
negative for
the
firms that

receive special attention in
40

NAFTA
, for example, communications, motor vehicles and equipment
,

and financial
41

services sectors;
h
owever,
t
he share
holders
gain if they invest
in
the

agricultural, textile
,

42

and apparel sectors.
In addition,

trade liberalizat
ion benefits both
value
–
added growth
43

firms and
labor
–
intensive

firms
in

the two neighboring countries.

Aggarwal et al.

(1
998)
44

study
the impacts of NAFTA on the values of the US firms under different
industry
45

characteristics a
nd the results show
that there
is

a
n

aggregate

positive impact of NAFTA
46

on equity prices

in the US
.
Industry
–
wise
,
they find

that
the
petroleum, auto products
,

47

and

telecommunications sectors
exhibit

significant
ly

negative returns while positive
48

returns
are observed
to
the
shareholders of food products, textiles, chemicals
,

and
49

machinery industries.

50

A
nother major work, Rodriguez (2003)
employs
a
n

event study to examine
51

investors’
perceptions
i
n

the manufacturing industries of
NAFTA’s
three participating
52

countries
. The findin
gs
argue

that

industry
–
wide
labor
–
capital

ratio is the most
important

53

determinant of
abnormal

returns
.

NAFTA would favor industries that use abundant
54

factors intensively. In a comprehensive study on regional trade agreements, Moser and
55

Rose (2014) examine 200
RTA
announcements
. They
find support
ive evidence

for
the
56

natural trading partner hypothesis that s
tock markets
react positively

when
the
countries
57

that are
existing
major trading partners sign the agreement and when poorer countries
58

collaborate with
wealthier ones but
find

no
evidence
showing
that
the
trade diversion
59

effects

are expected in the capital markets.

60

The studies
focusing
on the impact
s

of NAFTA
are not restricted

to examining
61

equity market returns for various countries.
Darrat and Zhong (2005)

look at the equity
62

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

3

of
20

market linkage
s among the

US
, Canada
,

and

Mexico
and

argue

that there is
an
intensified
63

equity linkage since NAFTA accord
s
.
Lóp
ez
–
Herrera

and Ortiz (2010)
then
provide

64

evidence of a time
–
varying integration process among NAFTA equity markets. In
65

another study, Chatterjee and Mitra (2000) analyze the effect
s

of NAFTA using a dummy
66

variable regression and find that the
influences
of the events

randomly appear

in
the
67

samp
le

countr
ies
.
It is evident based on
the
earlier works that the creation of NAFTA
is

a
68

critical

event that potentially
affects

stock
markets
in all three countries

of

North
69

America.
Therefore,
this paper
intends

to examine the chain of events starting wit
h
70

Election

Day and ending in the three countries reaching a new
trade
agreement, the
71

USMCA.

72

The rest of the paper
is organized

as follows
.

In
S
ection

2
,

we describe the event
73

chain and the hypotheses for the study. The empirical evidence
is provided

in
S
ection
3

74

followed by
the
conclusions in
S
ection
4
.

75

76

2
. Events and Hypotheses

77

Since the event period starts with the election and ends nearly two years later, we
78

only select

the most important events in
the study

period.
1

The event chain and its
79

hypothesized effect
s

on
the US,
Canadian,
and
Mexican stock indices
are
present
ed

in
80

Table 1.

81

1

The source of the information for most events was:
https://www.supplychaindive.com/news/NAFTA
–
timeline
–
ho
w
–
USMCA
–
happened/538663/

. We thank Supply Chain Drive and Mr. Edwin Lopez of Supply
Chain for permitting us to cite the information on their website.

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

4

of
20

Table 1.
Timeline of
the
Events in the Study

82

Event
Number

Date

Event Descriptions

Hypothesized
Effects on the US
Stock indices

Accept/R
eject

Hypothesized Effects on
Canadian and Mexican
Stock indices

Accept/Reject

Event 1

November 9,
2016

Donald John Trump was elected the 45th president of the United
States on Tuesday November 8
th
,

201
6
.
2

Positive

Accept

Negative

Reject for
Canada;

Accept
for Mexico

Event 2

January 23,
2017

President Trump to sign executive orders signaling TPP withdrawal,
NAFTA renegotiation. The long
–
awaited moves signal President
Trump's eagerness to renegotiate
trade deals, a move both Mexico and
Canada appear to favor.
3

Positive

Reject

Negative

Accept for
Canada;
Reject
for Mexico

Event 3

May 31,
2018

US moves forward with steel tariffs on EU, Canada, and Mexico.

The 25% steel and 10% aluminum tariffs on
imports from the
countries will go into effect at midnight tonight.

Negative

Accept

Negative

Reject

for

both
Canada

and
Mexico

June 1, 2018

US allies retaliate against ‘unacceptable’ steel tariffs.

Canada, Mexico
,

and the EU swiftly responded with retaliatory
measures on the US goods ranging from steel to felt
–
tipped pens.

Event 4

August 27,
2018

The US and Mexico reach
a
deal on NAFTA.

Canada is expected to return to the table promptly now the two parties
have resolved bilateral differences over issues including auto rules of
origin.

Positive

Reject

Negative for Canada and
Positive for Mexico

Accept
for

Canada

and
reject for
Mexico

August 28,
2018

What’s next for NAFTA?

A deluge of news emerged after the

US and Mexico reached a
bilateral deal yesterday, with speculation, Canada would be excluded
and NAFTA would be terminated.

August 31,
2018

US kicks off
the
process to sign NAFTA 2.0

The US has officially notified Congress of its intent to sign a
new
trade deal with Mexico
–

and maybe Canada
–

within 90 days.

2

https://www.nytimes.com/2016/11/09/us/politics/hillary
–
clinton
–
donald
–
trump
–
president.html

3

The source of the information for events
2
–
5 is:
https://www.supplychaindive.com/news/NAFTA
–
timeline
–
how
–
USMCA
–
happened/538663/

. We thank Supply Chain
Drive and Mr. Edwin Lopez of Supply Chain for permitti
ng us to cite the information on their website.

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

5

of
20

Event 5

September
30, 2018

NAFTA negotiators strike a new deal: USMCA

A new deal finally reached
legislators
’

desks last night.

Positive

reject

Positive

Accept

for
Canada

and

Mexico

83

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

6

of
20

The first event

on November 9, 2016,

is the surprise election of Mr. Trump as the
84

45
th

President of the United
States. During his election campaign, Mr. Trump mentioned
85

several times that he would withdraw the U
S

from

NAFTA should he be elected.
T
his
86

event
might

have
a
positive

impact on the returns of
the
US

stock indices, assuming
87

NAFTA
is considered

as a less favorable agreement to
the
US
, consistent with Mr.
88

Trump’s assertion. We also expect that both Mexico and Canada
would
have negative
89

returns since both countries
’

trade w
ith

the
US

would be
adversely impacted
. The
90

reasoning behind this argument

is that
any agreement that follows NAFTA would be less
91

favorable to the two neighboring countries of the U
S
.

92

The
second event
(
on
January 23, 2017)
i
s

the third day following the
swearing
–
in
93

of the new President when he signed the exec
utive orders to
renegotiate

NAFTA.
94

Consistent with the
first
event
that negative impacts are
hypothesize
d

for Mexican and
95

Canadian stock markets and positive returns for the
US

s
tock indices.

If the markets
96

perc
eive
the event
as

President Trump did, this would be a favorable event leading up to
97

comparative advantage to
US

companies as compared to their

Canadian
and Mexican
98

counterpart
ie
s.

99

The third major event occurred on May 31,
2018
,

and June 1, 2018.

On May 31,
100

201
8
,

the U
S

i
mposed steel tariffs on European Union, Canada
,

and

Mexico. EU,
101

Canada
,

and

Mexico

retaliated
on
US

goods the next day, June 1, 2
01
8. As tariffs
102

decrease profit margins in all countries, we
would
expect all returns to be negative for
103

markets of all three

countries

for
the
se

two day
s
.
This event

mark
s

a low point in the
104

NAFTA renegotiation.

105

The fourth event in the event chain occurred when
the
US

and Mexico

reached a
106

deal on August 27, 2018.
This

was followed by speculation the next day
(
August 28,
107

2018
)
, that Canada would
be excluded
.
On August 31,
2018
,

the US officially notified
108

Congress of its intent to sign a new deal with Mexico and maybe Canada within ninety
109

days.

This development is

then
hypothesized to produce positive returns for stock
110

indexes of Mexico and
the
US and negative returns for Canada over
the
five

days
.

111

The final event occurred when USMCA, the new agreement was announced late
112

on September 29, 2018
,

and we thus

consider the ne
xt day
,

September 30,
2018
,

to be the
113

event day. We hypothesize
that there
positive returns for all three North American
114

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

7

of
20

countries following the agreement as it
is expected

that the new deal is beneficial to all
115

parties.

116

117

3
. Empirical Evidence

118

The event
chain in the study starts on November 9, 2016
,

and ends nearly two
119

years later on September 30, 2018.
W
e
employ

daily percentage return
of

the indices
120

from January 14, 2014
,

through

January 22, 2019.
Such a
long period of data allows us
121

to have a long co
mparison period before the first event and three months of comparison
122

period after the last event.
The data

and the description of the indices

are

collected from

123

Bloomberg Database
.

T
he daily

percentage

returns on the indices
are

calculated
using
124

the
following method:

125

R
i
t

=
[
(
SI
t
–

SI
t
–
1
)/SI
t
–
1
]
*
100
,

(1)

126

where

R
it

is the return
of

the stock index for day t, and SI
t

is the daily closing
price

of the
127

index
on

day t, and SI
t
–
1

r
efers to the cl
osing
price
of the
index on day t
–
1
.

The sample

128

indices

include the major indices in

the
US, Canada
,

and Mexico.
4

These results are
129

presented
in the
next
sections
and Appendix
.

130

131

3.1

Analysis of
US Ind
ices

132

We first examine the returns
of
several

important

indices
in

the U
S
, the country
133

that initiated the revision of NAFTA and led the discussions for the subsequent
134

agreement USMCA.
We

exam the event chain

for the following major indices:

S&P 500
135

Index,
Dow Jones Industrial Average (DJI), Dow Jones Composite Average (DJC
OMP
)
,
136

NASDAQ Composite Index (NASDAQ), and Russell 2000 Index (RUSL2K).
The
137

dependent variable
in

the regression model is
daily returns for each of these indices

and
138

the model is

as

below:

139

R
it =

β
0

+
β
1
Event1 +
β
2
Event2 +
β
3
Event3 +
β
4
Event4 +
β
5
Event5

140

+
β
6

World Index +
β
7
LIBOR3M
,

(2)

141

4

The complete description of all the indices, and other control variables us
ed in the study are given
in

Appendix.

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

8

of
20

where
R
it

is the
daily
return
s

of

ind
ices mentioned

before
. The
first five
independent
142

variables are
dummy variables
representing
the event days
of

the
se

five events.
A

value
143

of 1
is set
for the event day(s), zero otherwise.
Two
control
variables
are
added

in the
144

regressions. The first
control variable
is the daily return
of

the
MSCI World Index
145

(MXWO)

and we utilize such index to control world stock market movements.

This

146

index

is a fr
ee
–
float weighted equity index
that
only

includes developed world markets.
147

To control for
interest rate movements on stock market indices,
the three
–
month LIBOR
148

(LIBOR3M)
rate

is used
.
The t
–
statistics reported in all the regressions are corrected
by

149

Newey and West (1987) heteroskedasticity and autocorrelation consistent covariance
150

matrix.

151

The results of the regressions for the US are
presented

in Table 2.

The

constant
152

terms are in
significant
.
The first event is the election of Mr. Trump as the President of the
153

United States. This event
generates

significant
impacts on

a
ll
five

indices.
All

f
ive

major
154

indices
have

positively significant abnormal returns
.

T
he hypothesis of Event 1 in the US
155

is

fully
supported.

156

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

9

of
20

Table 2: USA Results

157

The table below gives the results of the Regression Analysis for
US
–
based indices. The dependent variables in the regressions are the daily returns of The
158

Standard and Poor 500 (SP500),
Dow Jones Industrial Avera
ge (DJI), the Dow Jones Composite Average (DJCOMP), the NASDAQ Composite Index
159

(NASDAQ), and the Russell 2000 Index (RUSL2K). The independent variables are: C (Constant), the events (Events 1 to 5) the w
orld index (MXWO) and
160

the three month LIBOR rate. F
or all regressions, the Newey
–
West (1987) method was used to correct for autocorrelation and heteroskedasticity
.

F
or each
161

variable, the value of the regression coefficient, the standard error and the t
–
statistic value are given in that order below.

162

Index

SP500

DJI

DJCOMP

NASDAQ

RUSL2K

Intercept

0.01, 0.01, 1.47

0.01,
0.01
, 0.77

0.02, 0.02, 1.05

0.03, 0.02, 1.53

–
0.00, 0.03,
–
0.06

Event1

0.79, 0.33, 2.38
**

1.09, 0.01, 86.43***

0.11, 0.01, 9.06***

0.74, 0.01, 50.32***

2.76, 0.02, 171.65***

Event2

–
0.21, 0.33,
–
0.63

–
0.08, 0.01,
–
8.42***

–
0.37, 0.01,
–
36.96***

0.02, 0.01, 1.79*

–
0.21, 0.02,
–
13.27***

Event3

–
0.04, 0.24,
–
0.18

–
0.31, 0.09,
–
3.37***

–
0.39, 0.07,
–
5.27***

0.34, 0.06, 6.07***

–
0.25, 0.06,
–
4.20***

Event4

0.03, 0.15, 0.21

–
0.03,
0.03,
–
0.81

–
0.07, 0.03,
–
2.47***

0.23, 0.09, 2.48***

0.01, 0.20, 0.06

Event5

0.14, 0.33, 0.43

0.50, 0.02, 27.38***

0.24, 0.02, 12.59***

–
0.36, 0.03,
–
13.70***

–
1.63, 0.03,
–
53.06***

RetMXW
O

1.09, 0.01, 82.39
***

1.06, 0.04, 29.88***

1.00, 0.03,
30.54***

1.23, 0.04, 32.90***

1.19, 0.03, 35.54***

LIBOR3M

–
0.01, 0.01,
–
0.49

0.01, 0.01, 0.52

0.00, 0.01, 0.01

–
0.00, 0.02,
–
0.31

0.00, 0.02, 0.24

R
–
SQ.

0.84

0.80

0.77

0.75

0.66

N

1267

1267

1267

1267

1267

F
–
Statistic

972.57

708.71***

612.39***

525.79***

348.95***

Note:
*, **, and *** denote that the coefficient is statistically significant at the 10%, 5%
,

and 1% levels, respectively.

163

164

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

10

of
20

Event 2
occurs

when President Trump signed the executive orders to renegotiate
165

NAFTA. We hypothesize that this would be positive
information

for
the
US

stock
166

markets assuming that any new deal would favor the U
S

more
versus the
neighboring
167

countries. However,
in accordance with
the
returns of the
stock
indices
on

January 23
rd
,

168

2017,
three out of the five

major indices h
ave

significantly negative returns

and one
169

positive

(albeit NASDAQ
is
positive only at 0.10 le
vel). Overall, the decision to pull out
170

of NAFTA has
negatively
disturbed the markets

and
then rejects
the hypothesis
of Event
171

2.

172

Event 3 occur
s

during the period of
May
3
1, 2018
,

to

June 1, 2018
,

when the
US

173

after months of negotiations impose
s

steel tariffs on

the

EU, Canada
,

and Mexico
.

The
174

EU, Mexico
,

and Canada

retaliated
the next day
by imposing their tariffs on
US

goods.
175

The hypothesized effect for this two
–
day event for the
US

indices
is

negative since
the
176

tariffs reduce

the
profit marg
ins of companies and contribute to the erosion of shareholder
177

wealth. Three
out
of the
five
major indices
in

the
US

report significant
ly negative

returns
178

at 0.01 level
on

the
event
day
s
. Consistent with
the results of
Event 2,
the returns of
179

NASDAQ
index
are significant

and positive during the event period
.
Overall, the

results
180

suggest Event 3 is a negative event to the US

and support the hypothesis
.

181

Event 4 occur
s

from

August 27, 2018
,

and
ending
up to August 30, 2018.
182

Initially,
the US and
Mexico

reach a deal on NAFTA (as it was still called at the time),
183

with speculation that Canada might be excluded.
O
n August 30
, 22018

the U
S

noti
fies

184

Congress that it intended to sign a new deal with Mexico and possibly Canada shortly.
185

For the U
S
, we
might
expect this deal to have
a
positive
impact
,

and the results

do

show

186

t
hat

NASDAQ
has
significantly positive returns
to this event
, while DJCOMP
has
187

significantly negative returns.
T
he
SP500,
DJI
and
the RUSL2K
index
returns are
not
188

significant

during this
period
. This evidence

being mixed indicates

that

this event
did not
189

support the hypothesized positive impact
.

It also did not have a negative impact. We
190

reject the hypothesized positive impact since only one index has positive returns for the
191

day.

192

The f
inal event occur
s

when USMCA
was
announced

on September 30, 2018
.
193

USMCA
is
a replacement for NAFTA.
Such
an
event is

hypothesized to be a positive
194

one

for the stock markets since the new deal is expected to favor the U
S m
ore than
195

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

11

of
20

NAFTA
does
as the
President
’s intention
. The results show that
two of

the large
–
cap
196

indices

(
DJI and DJCOMP
)

have
positive returns
;

however,

the small
–
cap indices
197

(
NASDAQ and RUS
L
2K
)

have

negative returns.
We may conclude that

this event
198

generate
s

different
impact
s

on
these sectors

and is not
surely

a positive event
as
199

hypothesized
.

The hypothesis is only partially supported by the results.
In terms of
the
200

control variables,
the return of
the world index MXWO is positive and significant for all
201

regression

models, repr
esenting

there a

close relation
ship

between

the world index
and

all
202

major US indices
.
The coefficients of
LIBOR3M
are

in
significant in all

regressions
,

203

indicating that international interest rate movements during this time h
ave

no impact on
204

the returns
of

the major indices
in

the U
S
.

205

206

3.2
Analysis of
Canad
ian

Ind
ices

207

We
also
use Equation 2

to analyze the responses to the event

chain

in Canada
.
208

The stock market index is the
S&P/TSX
COMPOSITE INDEX (COMPOSITE
).

The
209

results are presented in Table 3
.

210

211

Table 3
: Canada Results

212

The table
shows

the results of the
r
egression
a
nalysis for Canada based indices.

The dependent variables in
213

the regressions are the daily returns of the S&P/TSX Composite Index (COMPOSITE).
The independent
214

variables are the 5 events,
the
return of the world index (MXWO), and the 3
–
month LIBOR rate.
F
or all
215

regressions, the Newey
–
West (1987) method was used to correct for autocorrelation and heteroskedasticity.
216

For each variable, the value of the regression coefficient, the standard error and the t
–
statistic value are given
217

in that order below.

218

Index

COMPOSITE

Intercept

–
0.00, 0.02,
–
0.11

Event1

0.49, 0.01, 33.20***

Event2

–
0.38, 0.01,
–
30.57***

Event3

–
0.18, 0.13,
–
1.32

Event4

–
0.21, 0.05,
–
4.36***

Event5

0.05, 0.02, 2.26**

RetMXW
O

0.75, 0.03, 25.44***

LIBOR3M

0.00, 0.01, 0.11

R
–
SQ.

0.57

N

1261

F
–
Statistic

241.49***

Note: see Table 2.

219

220

Table 3 show
s

that the constant

term

in the model
is

in
significant
and such
a result
221

is

consistent with
those in

the
US
.

For
Event 1, the election of Mr. Trump is
a
positive
222

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

12

of
20

event significantly positive returns
.
The hypothesis of
Even
t

1 in Canada
rejected.

Event
223

2
(
the announcement of the renegotiation of NAFTA by the President of the
US

on
224

January 23, 2017
)

is

a negative event for Canadian stock markets.
With
the coefficient
225

being negative
, the hypothesis of negative returns is

accepted
.
Event 3
occurs
on May 31,
226

2018
,

and the following day when the
US

imposed tariffs on Canada and Mexico and
227

these
two
countries retaliated the next day
is
a negative event for
the
Canadian stock
228

market.

Event 3
has no impact but we can reject the hypothesis of negative returns, as the
229

coefficient is significant.
Event 4 occur
s

between the period
of
August 27, 2018

through

230

August 30, 2018
,

and has
a hypothesized negative effect since
the
US

and Mexico had
231

reached a trade agreement and Canada was left out of the initial d
e
al. The Canadian
232

stock markets report negative returns for the COMPOSITE index
consistent with our
233

hypothesis
.
A positive impact generated by
Event 5

is
hypothesized.
T
he
COMPOSITE
234

index is positive and significant,
meaning
that
the USMCA is
seen as a positive
235

development
.

236

Consistent with the
US

indices,
the
MXWO world index
ha
s
a
significantly
237

positive coefficient
and

LIBOR3M
still
ha
s

no impact.
Canadian returns are more along
238

hypothesized lines with the exception of Event 1 and to some extent Event 3.

239

240

3
.
3

Analysis of
Mexico
Indices

241

For analyzing the Mexican stock market, we cho
o
se
four

major indices that are
242

viewed as
represent
ing

the country’s stock markets
, including

t
he S&P/BMV IPC (IPC),
243

S&P/BMV IPC SUSTAINABLE (SUST_IPC), FTSC BIVA RETURN INDEX
244

(FT
SE_BIVA
)

and
INMEX (largest stocks on S&P/BMV IPC
).

The regression models
245

are constructed as those of the US and Canada
,

and

R
it

d
enotes

the return
s

of these
four

246

indices
. For
the case of
Mexico, we use the
daily return of the
MSCI ACWI index
247

(RETMXWD)
which is a free
–
float weighted index instead of the
daily return of MXWO
248

used for
the U
S and Canada

since

the MSCI ACWI index includes both emerging and
249

world markets and is
, therefore,

the more appropriate measure as a control variable for
250

Mexico.

251

252

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

13

of
20

Table 4
: Mexico Results

253

The table
shows

the results of the
r
egression
a
nalysis for
Mexico

based indices.
The dependent variables in
254

the regressions are the daily returns of the
S&
P/BMV IPC (IPC), S&P/BMV IPC SUSTAINABLE
255

(SUST_IPC), FTSC BIVA RETURN INDEX (FTBIVA INDEX), and INMEX (largest stocks on S&P/BMV
256

IPC)
.

The independent variables are

the

5

events
,
the
return of

the world index (MXWO)
,

and the
3
–
month
257

LIBOR rate.
For all re
gressions, the Newey
–
West (1987) method was used to correct for autocorrelation and
258

heteroskedasticity.

For each variable, the value of the regression coefficient, the standard error and the t
–
259

statistic value are given in that order below.

260

Index

IPC

SUST_IPC

FTSE_BIVA

INMEX

Intercept

0.01, 0.03, 0.42

0.01, 0.03, 0.34

0.01, 0.03, 0.33

0.01, 0.03, 0.46

E
vent
1

–
2.20, 0.02,
–
126.29***

–
2.01, 0.02,
–
113.54***

–
2.00, 0.02,
–
118.64***

–
2.44, 0.02,
–
138.33***

E
vent
2

1.67, 0.02, 90.61***

1.54, 0.02,
82.51***

1.74, 0.02, 97.78***

1.76, 0.02, 94.86***

E
vent
3

0.16, 0.05, 3.15***

0.16, 0.06, 2.79***

0.20, 0.05, 4.35***

0.24, 0.06, 4.22***

E
vent
4

–
0.09, 0.12,
–
0.75

–
0.12, 0.13,
–
0.93

–
0.14, 0.12,
–
1.17

–
0.03, 0.13,
–
0.21

E
vent
5

0.60, 0.05, 12.07***

0.53, 0.05, 10.70***

0.63, 0.05, 13.07***

0.66, 0.05, 13.29***

R
etMXW
D

0.71, 0.03, 21.26***

0.73, 0.03, 21.62***

0.70, 0.03, 21.41***

0.75, 0.03, 22.18***

LIBOR3M

–
0.02, 0.03,
–
0.59

–
0.02, 0.03,
–
0.52

–
0.02, 0.03,
–
0.55

–
0.02, 0.03,
–
0.65

R
–
SQ.

0.36

0.36

0.36

0.38

N

1263

1263

1263

1263

F
–
Statistic

101.32***

102.10***

101.19***

108.76***

Note: see Table 2.

261

262

The results of the regressions are presented in Table 4. The constant
terms are

263

in
signif
icant in any of the regressions. Event 1, the election of Mr. Trump as President
,

264

le
a
d
s

to a significantly negative return in all four indices.
Such re
sults are

different from
265

the returns
of
the
US and Canada

but consistent with the hypothesized negative effect
.
266

The potential r
eason
is

that
,

during the election campaign, Mr. Trump had expressed the
267

opinion
s

that he would build a wall between the U
S

and Mexico. In addition to the threat
268

of pulling out of NAFT
A, which, as earlier studies have shown was a wealth
–
creating
269

event for Mexico. However,
the threat of pulling out of NAFTA
is possibly buffered

270

with an unrelated event of wall building for this event just
only
for Mexico.
For
Event 2
,

271

when the U
S

sugges
ted renegotiations
of NAFTA
. The returns
of

all
indices on this day
272

are all positive and significant
in

Mexico. This
result
s

in
exactly
the
opposite
of
the
273

hypothesized effect
.

A

reason could be that the negative returns had already occurred
274

after the e
lection and before the announcement of renegotiation

and thus the
275

announcement was only a formality and the stock markets had already priced in the
276

negative returns
.

The positive returns could indicate that the market feels that a new
277

agreement could now be negotiated.
We reject the hypothesis for Event 2.

278

Event 3
, the imposition of steel tariffs on trading partners by the
US

on May 31,
279

2018
,

followed by Mexico, Canada
,

and
the
EU retaliating the next day
is
hypothesized
280

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

14

of
20

as a negative event.
T
he four major indices

experience
positive returns on the event days.
281

The hypothesized effect not only did not happen but the exact opposite has happene
d
,
282

leading to a rejection of the hypothesis
.
Event 4
happens

when the
US

and Mexico
283

reached a deal on NAFTA that
Canada is not

initially

in
volve
d

but suggested that
284

negotiations are underway to include Canada within the next three months. This event
285

hypo
thesized as a positive event d
o
es
,

in fact,
ha
ve

no

impact

on the stock market of
286

Mexico.
None of the four indices show
s

any
significant

return
s

for Event 4.

Therefore,
287

Event 4

could

be
the least impactful event on the stock markets

and such results are
288

consistent with those in the US markets
.

Event 5
, the final event occurred on September
289

30, 2018
,

when
USMCA replaced NAFT
A. This event, which ended all uncertainty that
290

occurred for nearly two years
is
hypothesized to have a positive impact on the stock
291

markets
in
Mexico.
There are significantly

positive
for the four major indices
on the
292

event
day.
We accept our hypothesis for Event 5.
Consistent with the results in the US
293

and Canada
, the
coefficients
of the
world index RETMXWD
are also
positive and
294

significant
for
all eighteen indices. LIBOR3M has no significant coefficients in any of
295

the indices studied.

296

297

3.4. Robustness Checks

298

In this section, (G)ARCH models are then applied for robust checks of the

OLS
299

results. The ARCH process imposes an autoregressive structure on the conditional
300

variance which allows volatility shocks to persist over time. The stock markets usually
301

exhibit volatility clustering, implying large (small) changes are followed by larg
e (small)
302

changes, which has long been recognized as an important feature of stock return
303

behaviors. Before estimating the GARCH model, the important part to confirm the
304

existence of heteroscedasticity is on the ARCH test and the results (not presented her
e)
305

suggest that null hypothesis of no
h
eteroscedasticity/homoscedasticity is rejected at 1%
306

significance level for all stock markets and indices, indicating ARCH (1) effect are
307

present. To check the robustness of the ARCH results, this study further employ
s high
308

order ARCH effects (order 6 is used) and the results are still similar to those of ARCH (1)
309

test.

310

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

15

of
20

The table
presented in the appendix
reveals important results in identifying
311

whether these five events affect stock indices in the three countries. The

results show that
312

Event 1 (
the revision of NAFTA
) statistically significantly benefits the US and Canadian
313

stock markets but generates negative impacts on
the
Mexican stock market. For Event 2
314

(
the renegotiation of NAFTA
), we find there are positive impac
ts on Mexico but not for
315

the US and Canada. When
the US imposed steel tariffs on EU, Canada,
and

Mexico
316

(Event 3), most of the US and Canadian stock indices turn down (negative impacts) but
317

NASDAQ index rise
s
; Event 3 is also positive news to Mexico. In Au
gust 2018 (
the US
318

and Mexico

reach deal on NAFTA), Event 4 generally has no significant impacts on these
319

stock markets. Most of the indices for Canada and Mexico are negative but insignificant,
320

except for
COMPOSITE
. In the US, only NASDAQ index show
s

posit
ive responses. For
321

Event 5 (USMCA), this event brings great benefits to Mexico but not for Canada and the
322

small companies in the US since RUSL2K shows a negative coefficient. Generally, the
323

results of GARCH models are quite similar

to

those of OLS results
and the impacts of
324

these five events are convinced.

325

326

4
. Implications

and Conclusions

327

The
paper
examines the impact
s

of USMCA

on the stock markets of the US,
328

Canada, and Mexico
, starting with the
US

Presidential Election of 2016, using a chain of
329

major events.
Event 1
, the election of President Trump
fairly generates

positive returns
330

for
the
US

and Canada
but

brings
negative returns
to

Mexico. Based on the hypothesis,
331

we cannot reject the positive
responses

from
the
US
.
Canada
’s COMPOSITE

index
332

contrary to
the
hypothesis had
positive

returns
. Consistent with the hypothesis,

negative
333

returns

accrued for all
Mexican indices. Evidence
implies
that the
stock
markets
in
334

Canada d
id

not
consider

the election of Mr. Trump as a negative event. The
reason
is that

335

the
US

and Canadian stock
markets

are highly correlated

with each other. This strong
336

positive response from the event in the US market is probably reflected in the Canadian
337

stock index al
so.

338

Event 2
(
when President Trump ordered NAFTA renegotiation
)

i
s hypothesized
339

as a negative event for Mexico and Canada and
a
positive
event
for
the
US
. The results
340

suggest that the
event
is
negative
to

the
US

and Canada and positive for Mexico.
341

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

16

of
20

T
herefore,

the hypothesis
is supportive
for Canada
. For
the
US

and Mexico

stock
342

markets
,
the
hypothesis
is rejected
.
For
the
US
, the reason for this negative returns could
343

be the short
–
term implications of renegotiating a deal would mean lower stockholder

344

returns, albeit temporarily. A second reason could be that

the
US

h
as

traditionally
345

supported free trade among nations. For Mexico, it is possible that the positive returns
346

stem from the fact that a long
–
anticipated event has finally occurred.
It could

now start
347

renegotiating a deal
,

and the picture has become clearer.

It could also be that Mexico
has
348

negative returns for Event 1
,

which
is
based on the pre
–
election speeches of the incoming
349

President covered for the impact of
the
removal of NAFTA.

It is possible that the second
350

event was well anticipated in Mexico and the actual announcement was already priced
351

into the stock market.

352

Event 3, the announcement of steel tariffs by
US

and retaliation by Canada and
353

Mexico
is

hypothesized as negative
n
ews

for
the

stock markets

in these three countries
.
354

The results suggest that the negative returns
are observed

in
the
US, no impact in Canada
355

and positive returns in Mexico.

We do not reject our hypothesis for Event 3

for Canada
356

and Mexico
.

For
Event 4
(
when Mexico and
the US

reached a deal and speculation arose
357

that Canada might be left out a new agreement
),

t
he US notified Congress of its intention
358

to sign a new deal with Mexico and possibly
involve
Canada.
In the
US
, the hypothesis
359

is rejected as
three of the five indices do now show any impact while the other two are
360

mixed.

As
expected
, Canada ha
s

negative returns

so we accept the hypothesis for that
361

country, while we reject the hypothesis based on no observed impact in Mexico.

Event 5
362

(
when US
MCA was announced on September 30, 2018
),

it appears that the
US

stock
363

markets
do not

consider

this
news is

a positive event.
Two indices had positive returns
364

while two others have negative returns. Canada has positive returns and so does Mexico.
365

The res
ults of this
specific

event are comparable to some earlier studies on the signing of
366

NAFTA. As Hanson and Song (1998)
indicate
,

the
US firms received insignificant
367

returns while Mexican firms received positive returns at the announcement of NAFTA in
368

1993.

A second reason could be as mentioned by Moser and Rose (2014)
5

who mention
369

that stock markets tend to
incr
e
ase

more in poorer countries signing regional trade
370

5

See Page 38 of Moser, C. and Rose, A.K., 2014. Who benefits from regional trade agreements? The view
from the stock market.
European Economic Review
,
68
, pp.

31
–
47.

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

17

of
20

agreements than richer ones. In
the
case of the USMCA, Canada and
the US

are
371

wealthier

than Mexico.

372

The process of globalization has enabled investors to invest in financial markets
373

all over the world. However, the appearance of global investors has tightened
374

relationships between financial markets in different parts of the world. This, in
turn, has
375

made international portfolio diversification a very difficult task. A deeper analysis of the
376

existence and strength of relationships between markets for risk management and optimal
377

portfolio allocation has become important.
The study
brings sever
al important
378

implications

for
academics and practitioners. For academics, the study shows the impact
s

379

of a long chain of events
resulting
from the abolishment of one trade agreement and its
380

replacement

with

another on the stock markets of the trading countries. The study also
381

presents
how it is possible for a country to pull out of an existing agreement and lead
a
382

discussion on its replacement and yet see the major stock indices not showing immediate
383

positi
ve return
s
.
For practitioners,
it is important that understanding

pulling out of trade
384

deals does not guarantee that the next deal would be better
for

stock market performance
s
385

but might favor the stock markets in

developing countries.
In addition, the r
esults of
386

Event 3
show
that tariffs would lead to negative returns for all parties

and

the
387

relationship
s

between Mexico and the other two stock markets
are

not as strong as the
388

relationship between the US and Canada
.

Investors

could use this information t
o go long
389

or short
positions
on index
–
based ETFs of the three
sample
countries
when
similar events
390

happen in the future.

391

392

A
PPENDIX OF THE STOCK INDICES

393

US
A

394

395

1.

S&P 500 (SP500):
The S&P 500 is the stock market index that
measures the
396

performance of 500 large companies listed on the stock exchanges of the United States.
397

It is one of the leading indices used by portfolio managers for representing the U.S.
398

stock market.

399

2.

Dow Jones Industrial Average (DJI)
:
The
Dow Jones Indu
strial Average is
400

compiled by Dow Jones as a way to gauge the performance of the industrial
401

component of America's stock markets. It is the oldest continuing U.S market index.

402

3.

Dow Jones Composite Average (DJCOMP)
:
The Dow Jones Composite Average is
403

a pric
e
–
weighted average that tracks 65
US

prominent companies. The average is a
404

combination of the Dow Jones Industrial, Transportation, and Utilities Averages.

405

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

18

of
20

4.

NASDAQ Composite Index (NASDAQ)
:
The NASDAQ Composite Index is a
406

broad
–
based capitalization
–
weighted

index of stocks in all three NASDAQ tiers:
407

Global Select, Global Market
,

and Capital Market. The index was developed with a
408

base level of 100 as of February 5, 1971.

409

5.

Russell 2000 Index (RUSL2K)
:
The Russell 2000 Index is comprised of the smallest
410

2000 com
panies in the Russell 3000 Index, representing approximately 8% of the
411

Russell 3000 total market capitalization. The real
–
time value is calculated with a base
412

value of 135.00 as of December 31, 1986. The end
–
of
–
day value is calculated with a
413

base value of
100.00 as of December 29, 1978.

414

415

CANADA

416

417

1.

S&P/TSX Composite Index (
COMPOSITE
)
:
The S&P/Toronto Stock Exchange
418

Composite Index is a capitalization
–
weighted index designed to measure market
419

activity of stocks listed on the TSX. The index was developed with a
base level of
420

1000 as of 1975. The sectors available under SPTSX Index GRPS<GO> do not price
421

intraday. This index contains investment trusts effective 12/19/05. For the S&P/TSX
422

Equi
ty Index please see TXEQ Index.

423

424

MEXICO

425

426

1.

S&P/BMV IPC (
IPC
)
:
The S&P/BMV IPC

seeks to measure the performance of the
427

largest and most liquid stocks listed on the Bolsa Mexicana de Valores. The index is
428

designed to provide a broad, representative, yet easily replicable index covering the
429

Mexican equities market. The constituents ar
e weighted by
the
modified market cap
430

subject to diversification requirements.

431

2.

S&P/BMV IPC SUSTAINABLE (
SUST_IPC
)
:
The S&
P/BMV IPC Sustainable is
432

designed to measure the performance of Mexico's leading companies in terms of
433

economic, environmental, and social criteria, providing an objective benchmark for
434

managing sustainability investment portfolios. The constituents are we
ighted by
the
435

modified market cap subject to diversification requirements.

436

3.

FTSE BIVA PRICE RETURN INDEX (FTBIVA INDEX)
:
T
he FTSE BIVA
437

Index is designed to represent the behavior of Mexico's stock market. Using the FTSE
438

Global Equity Index Series (
T
GEIS) as

a base, the FTSE BIVA Index uses a higher
439

liquidity threshold to reflect the performance of liquid Mexican companies. The FTSE
440

BIVA Index also includes REITs.

441

4.

S&P/BMV INMEX INDEX (INMEX)
:
The S&P/BMV INMEX index is designed
442

to measure the performance of t
he 20 largest and most liquid stocks of the S&P/BMV
443

IPC. Index constituents are weighted by modified market cap, subject to
444

diversification requirements.

445

446

WORLD INDEXES AND LIBOR (CONTROL VARIABLES)

447

448

1.

MSCI World Index
:
The MSCI World Index is a free
–
float
weighted equity index. It
449

was developed with a base value of 100 as of December 31, 1969. MXWO includes
450

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

19

of
20

developed world markets and does not include emerging markets. MXWD includes
451

both emerging and developed markets.

452

2.

MSCI ACWI Index
:
The MSCI ACWI Index i
s a free
–
float weighted equity index.
453

It was developed with a base value of 100 as of December 31
,

1987. MXWD includes
454

both emerging and developed world markets. For developed markets only, please see
455

MXWO.

456

3.

ICE LIBOR USD 3 Month
:
London
–

Interbank Offered

Rate
–

ICE Benchmark
457

Administration Fixing for US Dollar. The fixing is conducted each day at 11am &
458

released at 11.45am (London time). The rate is an average derived from the quotations
459

provided by the banks determined by the ICE Benchmark Administration
.

460

461

462

463

Appendix Table. Robustness Checks of GARCH (1,1) Models

464

This table shows the results of
the
GARCH (1,1) models. Due to the limited space, we only show the
465

estimated coefficients and T
–
Stat values for the dependent variables in mean equations. Other results,
466

including the coefficients of variance equations are available upon request.
GARCH (1,1) m
odel applied in
467

the paper can be expressed as:

468

Mean Equation:


=

+
∑


×






+

×
(

)
+

×

3

+


,


/
469

Φ



~
N
(
0
,
ℎ

)
,

470

Variance Equation:
ℎ

=

+

×





+

×
ℎ



,

471

where
ω
er
,
α
≥
0
, and
β
≥
0
.
Φ

is the information set of all information through time
t
−
1
.
The
472

GARCH(1,1) is weakly stationary if
α
he

G
,

is the news about volatility from the previous period (the
473

ARCH term), and
ℎ

represents the conditional variance and is the last period forecast variance (the
474

GARCH term) and must be nonnegative.
In addition, we also tried GARCH (2,1), GARCH (1,2), and
475

EGARCH (1,1) and the results are similar
to

those of GARCH (1,1) models.

476

Panel
A

US: SP500

US: DJI

US: DJCOMP

US: NASDAQ

US: RUSL2K

Coeff
.

T
–
Stat

Coeff.

T
–
Stat

Coeff
.

T
–
Stat

Coeff
.

T
–
Stat

Coeff
.

T
–
Stat

Constant

0.016
1

2.3150

**

2.3150

1.7941

**

0.015
1

1.5890

0.027
4

2.0856

**

0.004
2

0.2643

Event1

0.798
9

2.9125

**
*

2.9125

3.9201

**
*

0.123
9

2.2305

**

0.743
2

1.6740

*

2.011
9

2.1578

**

Event2

–
0.208
0

–
1.1276

–
1.1276

–
1.6575

*

–
0.378
6

–
1.9161

*

0.021
4

2.0552

**

–
0.227
1

–
0.3612

Event3

–
0.021
2

–
0.1216

–
0.1216

–
2.3896

**

–
0.385
9

–
1.7205

*

0.339
7

1.7441

*

–
0.240
0

–
0.5670

Event4

0.020
8

0.2218

0.2218

–
0.1399

–
0.098
2

–
1.6460

*

0.216
2

2.1556

**
*

0.011
4

0.0416

Event5

0.152
4

0.6015

0.6015

2.5937

**
*

0.250
0

3.7001

**
*

–
0.364
5

–
1.6443

–
1.620
0

–
2.5870

**
*

RetMXW
O

1.030
0

74.713
6

**
*

74.713
6

58.780
9

**
*

0.972
7

55.375
7

**
*

1.195
9

50.831
2

**
*

1.187
4

49.148
7

**
*

LIBOR3
M

0.009
3

0.9314

0.9314

2.2400

**

–
0.003
0

–
0.2368

0.009
6

0.5296

–
0.011
3

–
0.5072

477

J. Risk Financial Manag.
20
20
,
1
3
, x FOR PEER REVIEW

20

of
20

Panel B

Canada:
COMPOSITE

Mexico: IPC

Mexico:
SUST_IPC

Mexico:
FTSE_BIVA

Mexico:
INMEX

Coeff.

T
–
Stat

Coeff
.

T
–
Stat

Coeff.

T
–
Stat

Coeff.

T
–
Stat

Coeff.

T
–
Stat

Constant

–
0.011
1

–
0.229
6

0.014
0

0.793
8

0.013
6

0.766
7

0.0126

0.7378

0.012
7

0.714
5

Event1

2.262
1

1.843
9

*

–
4.416
8

–
7.829
9

***

–
0.096
4

–
2.198
9

**

–
0.2316

–
2.4167

**

–
0.237
9

–
2.484
9

**

Event2

2.017
9

1.053
8

1.617
5

1.932
6

*

1.486
3

1.939
6

*

1.6985

1.0788

1.715
5

1.726
9

*

Event3

–
0.202
3

–
1.289
8

0.141
9

2.350
5

**

0.147
8

1.748
4

*

0.1899

2.4919

**

0.213
0

2.527
3

**

Event4

–
0.653
4

–
4.802
6

***

–
0.132
1

–
0.453
5

–
0.157
2

–
0.521
8

–
0.1808

–
0.6302

–
0.070
5

–
0.220
4

Event5

–
0.139
1

–
1
.
6
47
1

*

0.560
6

1.893
9

**

0.498
3

1.727
5

*

0.5950

2.9892

**
*

0.615
7

1.829
3

*

Ret
MXW
D
#

0.472
1

7.160
7

***

0.699
2

28.06
35

***

0.716
7

26.62
96

**
*

0.6793

28.739
8

**
*

0.740
2

28.21
73

**
*

LIBOR3
M

–
0.049
4

–
0.633
8

–
0.057
7

–
2.446
3

**

–
0.059
3

–
2.503
8

**

–
0.0599

–
2.6627

**
*

–
0.060
8

–
2.603
4

**
*

#

F
or Canadian and Mexican stock markets, we use MXWO and MXWD respectively.

*, **, and *** denote
478

that the coefficient is statistically significant at the 10%, 5%
,

and 1% levels, respectively.

479

480

481

R
eferences

482

Aggarwal, R., Long, M., Moore, S. and Ervin, D., 1998. Industry Differences in NAFTA
’
s
483

Impact on the Valuation of US Companies.

International Review of Financial
484

Analysis
,

7
(2), pp.137
–
152.

485

Chatterjee, A. and Mitra, A., 2000. Rate of Return of Stock Index and the NAFTA Process:
486

An Analysis of North America and ASEAN Financial Markets.
Managerial Finance
,
487

26
(12), pp.1
–
12.

488

Darrat, A.F. and Zhong, M., 2005. Equity market linkage and multinationa
l trade accords:
489

The case of NAFTA.
Journal of International Money and Finance
,
24
(5), pp.793
–
817.

490

Hanson, R.C. and Song, M.H., 1998. Shareholder wealth effects of free trade: US and
491

Mexican stock market response to NAFTA.

International Review of Economics

&
492

Finance
,

7
(2), pp.209
–
224.

493

López
–
Herrera, F. and Ortiz, E., 2010. Cointegration trends among the NAFTA equity
494

markets.
Revista de Economía Mundial

26.

495

Moser, C. and Rose, A.K., 2014. Who benefits from regional trade agreements? The view
496

from the stock m
arket.
European Economic Review
,
68
, pp.31
–
47.

497

Newey, W.K., and West, K.D., 1987. A Simple, Positive Semi
–
Definite, Heteroskedasticity
498

and
Autocorr
el
ation
Consistent Covariance Matrix
. Econometrica,
55(3), 703
–
08.

499

Rodriguez, P., 2003. Investor expectations and the North American free trade agreement.
500

Review of International Economics
,
11
(1), pp.206
–
218.

501

502