- 1 The authors would like to thank three anonymous referees for their helpful comments. Any remaining (...)
1By the time Mexican authorities began considering free trade with the U.S. and Canada, the country had already advanced substantially towards trade liberalization (Weintraub 1991, Gasca Zamora 2006). The possible consequences of increased trade liberalization on regional economic inequalities had however not attracted much attention in Mexico (Esquivel 2014). Ten years after NAFTA came into force it had become obvious that the agreement was exacerbating such inequalities, in particular by concentrating most of its economic benefits in urban areas of Central, Central-Western and above all, Northern Mexico (Gasca Zamora 2006). By 2014, NAFTA at 20 had become a hot topic. The agreement’ legacy has recently been addressed in several publications (Villarreal and Fergusson 2013, U.S. Chamber of Commerce 2014, Beachy 2014), a special issue of Foreign Affairs (January/February 2014) and a number of events and conferences organized in both the U.S. and Mexico.
2Under NAFTA, the trilateral U.S.-Mexico-Canada trade relationship has consisted to a large extent of two bi-national relationships: U.S.-Mexico and U.S.-Canada. Regional aspects of U.S.-Canada trade have been widely studied (McCallum 1995, Helliwell and McCallum 1995, Helliwell 2002, Globerman and Storer 2013). This body of research has shown the importance of the border effect in shaping regional trade patterns between the two countries.
3In contrast, there exists a paucity of information on the regional effects of trade within Mexico (Aguilar Barajas 2008, Esquivel 2014, Garduño Rivera 2014). The regional analysis published so far has focused strictly on the effects on U.S. states of trade with Mexico as a whole (Wilson 2011, U.S. Chamber of Commerce 2014). This is unfortunate in light of the significance of trade for regional economic development in Mexico.
4This paper addresses several regional aspects of U.S.-Mexico trade, including the impact of NAFTA on cross-border regional economic integration. The rest of the paper is organized as follows. The first section offers a brief overview of the U.S.-Mexico border region as well as an aggregate picture of Mexico’s trade with the U.S. and Canada. The second section analyzes state-level U.S.-Mexico trade patterns. The third section highlights the role of the Laredo Port of Entry as the heart of bi-national trade. The fourth section quantifies the impact of NAFTA on economic integration as well as current levels of economic linkages in the Texas-Northeastern Mexico region. Finally, section five concludes.
5Table 1 reports population and population growth for the U.S.-Mexico Border Region. Three distinct definitions of this region are employed. The first definition, known as the Border Strip, includes the forty Mexican municipalities (municipios) and twenty-four U.S. counties located along the international border. The second definition, known as the Border Area, includes 300 kilometers of Mexican territory south of the international border, and 100 kilometers of U.S. territory north of the international border. This definition has been employed by the North American Development Bank (NADBank) and the Border Environment Cooperation Commission (BECC) for programmatic purposes since 2004 and captures important cross-border processes, for example economic corridors connecting urban markets such as Hermosillo-Tucson and Saltillo-Monterrey-Laredo. The third definition, known as the Transborder Region, includes the ten Border States: Texas, New Mexico, Arizona and California in the U.S, and Tamaulipas, Nuevo León, Coahuila, Sonora, and Baja California in Mexico. The Border States aggregate almost 91 million inhabitants, a large and significant number in and of itself that represents more than a fifth of the two countries’ combined population. Moreover, population within the border region has been growing at a much faster pace than in the rest of the two countries.
Table 1. U.S.-Mexico Border Region: Population (2010) and Annual Population Growth (2000-2010)
|
Population
(2010)
|
Average annual growth (%) (2000-2010)
|
Border Strip
|
|
|
Mexico, border municipios
|
7,304,901
|
2.24
|
United States, border counties
|
7,303,754
|
1.62
|
Total
|
14,608,655
|
|
|
|
|
Border Area
|
|
|
Mexico, 300 km
|
17,048,419
|
2.04
|
U.S., 100 km
|
13,967,038
|
2.28
|
Total
|
31,015,457
|
|
|
|
|
Transborder Region
|
|
|
Mexico, border states
|
19,894,418
|
1.95
|
United States, border states
|
70,850,713
|
1.49
|
Total
|
90,745,131
|
|
|
|
|
Mexico
|
112,336,538
|
1.52
|
United States
|
308,745,538
|
0.97
|
Source: Authors’ own, with data from Wilson and Lee (2013: 29).
6Map 1 includes the three definitions of the border region presented earlier: Border Strip, Border Area and Transborder Region. The map also identifies some specific non-border states in the U.S. as well as in Mexico. As will be shown later, these non-border states have accounted for a significant share of overall U.S.-Mexico trade. Finally, the map also reports the location of major urban areas within the Transborder Region and the aforementioned non-border states.
Map. The U.S.-Mexico Border Region
Source: Authors’ own.
7Figure 1 illustrates how the total value of goods (imports plus exports) traded by Mexico with its two NAFTA partners has grown between from 1994 to 2012; Table 2 presents the data. Overall, the value of that trade has risen by over 400% over that period. In 1994, the U.S. accounted for 96% of Mexico’s NAFTA trade. Although Mexico’ trade with Canada has increased at a faster pace than with the U.S. (at an average yearly growth rate of 12.6% for the former, versus 9.8% for the latter) at the end of the period the U.S. still largely dominated Mexico’s NAFTA trade with a share of 94%.
Figure 1. Value of Mexico’s NAFTA trade in goods (imports plus exports), by country of origin-destination, 1994-2012 (millions of nominal U.S. dollars)
Source: Authors’ own, with data from Dawson et al. (2013: 40).
Table 2. Value of Mexico’s NAFTA trade in goods (imports, exports and total), by country of origin-destination, 1994-2012 (millions of U.S. dollars)
|
Mexico NAFTA Exports (1)
|
Mexico NAFTA Imports (2)
|
Total(1)+(2)
|
Year
|
U.S.
|
Canada
|
Total
|
U.S.
|
Canada
|
Total
|
|
1994
|
49,494
|
3,313
|
52,807
|
50,843
|
793
|
51,636
|
104,443
|
1995
|
62,101
|
3,899
|
66,000
|
46,311
|
846
|
47,157
|
113,157
|
1996
|
74,297
|
4,426
|
78,723
|
56,792
|
923
|
57,715
|
136,438
|
1997
|
85,938
|
5,072
|
91,010
|
71,388
|
922
|
72,310
|
163,320
|
1998
|
94,629
|
5,180
|
99,809
|
78,772
|
989
|
79,761
|
179,570
|
1999
|
109,721
|
6,418
|
116,139
|
86,909
|
1,085
|
87,994
|
204,133
|
2000
|
135,926
|
8,120
|
144,046
|
111,349
|
1,370
|
112,719
|
256,765
|
2001
|
131,338
|
7,829
|
139,167
|
101,296
|
1,779
|
103,075
|
242,242
|
2002
|
134,616
|
8,115
|
142,731
|
97,470
|
1,541
|
99,011
|
241,742
|
2003
|
138,060
|
8,698
|
146,758
|
97,412
|
1,578
|
98,990
|
245,748
|
2004
|
155,902
|
10,322
|
166,224
|
110,835
|
2,379
|
113,214
|
279,438
|
2005
|
170,109
|
12,047
|
182,156
|
120,365
|
2,778
|
123,143
|
305,299
|
2006
|
198,253
|
14,125
|
212,378
|
133,979
|
3,858
|
137,837
|
350,216
|
2007
|
210,714
|
15,984
|
226,698
|
136,092
|
4,613
|
140,705
|
367,403
|
2008
|
215,942
|
16,807
|
232,749
|
151,220
|
5,482
|
156,702
|
389,451
|
2009
|
176,654
|
14,480
|
191,134
|
128,892
|
4,206
|
133,098
|
324,233
|
2010
|
229,908
|
21,470
|
251,378
|
163,473
|
4,863
|
168,336
|
419,714
|
2011
|
263,106
|
24,842
|
287,948
|
197,544
|
5,537
|
203,081
|
491,029
|
2012
|
277,653
|
25,534
|
303,187
|
216,331
|
5,391
|
221,722
|
524,909
|
Source: Authors’ own, with data from Dawson et al. (2013: 40).
8This section focuses on surface trade (i.e. by truck and railway) of goods between the U.S. and Mexico; on average more than 80% of yearly U.S.-Mexico trade is transported by land (TDOT 2013). We use data from the U.S. Bureau of Transportation Statistics (BTS) that link the origin of an export in a U.S. state with a destination state in Mexico. Normal caveats with this type of data apply here. The data identify either the origin of a shipment’s movement, or the location of the exporter. Rather than the actual state where the export is generated, data may therefore refer to the location of a broker, wholesaler or distributor. However, it is believed that overall the data do reflect the real patterns of trade between the two countries. Furthermore, Mexico does not disaggregate its own exports to the U.S. by state of origin; only the total exports from Mexico to a given state in the U.S. are known.
9In 2012 the value of U.S.-Mexico trade reached over 390 billion dollars, up from 97 billion in 1995. Table 3 presents trade data for four selected years (1995, 2000, 2005 and 2012). In each case, the table reports the top five U.S. states according to the total value of their trade with Mexico. Clearly, U.S.-Mexico trade has largely concentrated in a few states. Texas, California and Michigan have remained the three top states throughout the period; the top five states have consistently accounted for two thirds or more of total trade; and Texas, by far the most important trading state, has kept its share fairly constant at about a third of the total.
Table 3. U.S.-Mexico surface trade in goods (imports and exports) and top 5 U.S. states, millions of current U.S. dollars, 1995, 2000, 2005, 2012
State
|
1995
|
2000
|
2005
|
2012
|
United States
|
96,711
|
210,595
|
239,677
|
391,407
|
Texas
|
32,982
|
68,956
|
77,943
|
144,344
|
California
|
15,339
|
32,122
|
41,476
|
56,526
|
Michigan
|
12,657
|
22,794
|
23,635
|
48,106
|
Arizona
|
5,200
|
10,452
|
8,446
|
12,808
|
Tennessee
|
2,620
|
|
|
|
Indiana
|
|
6,053
|
|
|
Illinois
|
|
|
8,057
|
13,518
|
Share of Top 5
|
71.1%
|
66.7%
|
66.6%
|
70.3%
|
Share of Texas
|
34.1%
|
32.7%
|
32.5%
|
36.9%
|
Source: Authors’ own, with data from the U.S. Department of Transportation Research and Innovative Technology Administration, Bureau of Transportation Statistics.
10Table 3 also shows the dominance of Border States in the U.S.-Mexico trade relationship: in every year of observation, we find three Transborder Region states (Texas, California and Arizona) in the list of the top five states. Location within that region however is not sufficient for establishing significant trade ties with Mexico, as exemplified by New Mexico which ranked 35th in 2012 with only a billion dollars’ worth of trade. Neither is location along the border a necessary condition for significant trade with Mexico: Michigan, a distant non-border state, has held the third position throughout 1995-2012; in 2012 it traded with Mexico 3.5 times more than Arizona and almost as much as California. This is explained by Michigan’s deep historical ties to the U.S. auto industry, which in the last decades has invested heavily in the Mexican automotive sector.
11In addition to geographical proximity, the size of a state’s economy obviously matters. In this sense, New Mexico’s limited trade ties with Mexico reflect its relatively small economic size compared to Texas, California or Michigan. For the same reason, in Mexico the Federal District and the State of Mexico (non-border locations) are relatively large in economic terms and therefore their strong trade links with the U.S. (which are documented later on in this section) should not cause any surprise.
12Focusing on exports (or imports) does not change the patterns identified previously in the trade data. Table 4 presents export data for 2012 for the top eight U.S. states according to the total value of their exports to Mexico. The data show that the top five exporters are the same as the top five traders of the previous table. These five states account for 70% of total U.S. exports to Mexico, the three trans-border region states alone obtain a share of over 60% and Texas is by far the largest single exporter. Data not shown here reveal that these features have largely remained identical throughout the 1995-2012 period.
Table 4. U.S.-Mexico surface exports of goods and top 8 U.S. states, millions of current U.S. dollars, 2012
State
|
Value of Exports
|
United States
|
171,878
|
Texas
|
75,964
|
California
|
22,697
|
Michigan
|
10,316
|
Arizona
|
6,120
|
Illinois
|
5,937
|
Ohio
|
4,394
|
Tennessee
|
3,986
|
Indiana
|
3,714
|
Share of top 5 states
|
70.4%
|
Share of Transborder states
|
61.0%
|
Share of Texas
|
44.2%
|
Source: Authors’ own, with data from the U.S. Department of Transportation Research and Innovative Technology Administration, Bureau of Transportation Statistics.
13The importance of the border region in U.S.-Mexico trade can also be observed on the Mexican side of the border. Table 5 presents, for the top three U.S. exporting states of Table 4 (Texas, California and Michigan) a breakdown of their exports by recipient Mexican state (only the top ten recipients are listed). Mexican Border States account for 59% of Texas exports to Mexico, 75% of California’s and 36% of Michigan’s. In the case of Texas, exports to Mexican Border States are almost entirely concentrated in the four Northeastern states (Tamaulipas, Nuevo León, Coahuila and Chihuahua). Note that for both Texas and California, the main Mexican recipient is a state just across the border (Chihuahua and Baja California, respectively).
14Notwithstanding the preceding, the data show once again that trade reaches far beyond the border region. For all three U.S. states, the Federal District and State of Mexico (the Mexican industrial heartland) are among their top four Mexican clients. Both are especially important for Michigan as they represent 45% of all its exports to Mexico.
15The regional features of U.S.-Mexico trade identified here point to varied, multi-layered border and trans-border dynamics. These dynamics feature a notable border effect but, obviously said, this effect only captures part of the whole picture.
Table 5. Exports from top three trading U.S. states to top ten recipient Mexican states, millions of current U.S. dollars, 2012
|
Texas
|
California
|
Michigan
|
|
Total
|
75,964
|
Total
|
22,697
|
Total
|
10,316
|
1
|
Chihuahua
|
24,200
|
Baja California
|
12,102
|
Distrito Federal
|
3,307
|
2
|
Distrito Federal
|
11,839
|
Chihuahua
|
3,351
|
Coahuila
|
1,644
|
3
|
Tamaulipas
|
9,509
|
Distrito Federal
|
2,156
|
Estado de México
|
1,355
|
4
|
Estado de México
|
8,509
|
Estado de México
|
1,931
|
Nuevo León
|
807
|
5
|
Coahuila
|
6,420
|
Sonora
|
829
|
Chihuahua
|
797
|
6
|
Nuevo León
|
4,665
|
Jalisco
|
693
|
Guanajuato
|
496
|
7
|
Guanajuato
|
3,409
|
Nuevo León
|
469
|
Tamaulipas
|
443
|
8
|
Jalisco
|
2,289
|
Baja California Sur
|
231
|
Querétaro
|
346
|
9
|
Querétaro
|
721
|
Sinaloa
|
156
|
Puebla
|
332
|
10
|
San Luis Potosí
|
666
|
Querétaro
|
132
|
Jalisco
|
117
|
|
Share of border states
|
59.0%
|
|
74.8%
|
|
35.8%
|
Source: Authors’ own, based on data from the U.S. Department of Transportation Research and Innovative Technology Administration, Bureau of Transportation Statistics.
16Laredo in Texas is one of the most important Port of Entry (POE) for all U.S. trade. In terms of the value of trade (imports plus exports), Laredo ranked sixth in the U.S. with 163 billion dollars in 2012. The first position was held by the Los Angeles POE, which handled 217 billion dollars of trade. As shown earlier, Texas accounts for an important share of U.S.-Mexico surface trade in goods. Laredo is the major POE in Texas; it is also the major POE between the U.S. and Mexico, and plays a leading role as a NAFTA POE. Table 6 shows how Laredo has increased its relevance as a NAFTA port of entry, to become number one in 2012. It is also worth noting that six of the ten top trade ports of entry between Mexico and the U.S. are located in Texas.
Table 6. Top U.S. Ports of Entry (POE) for surface trade with Canada and Mexico, 1995-2012 (volume of operations in billions of nominal U.S. dollars)
|
1995
|
2000
|
2005
|
2012
|
U.S. trade with Canada and México
|
345
|
546
|
659
|
910
|
10 Top POE
|
238
|
407
|
481
|
654
|
Laredo, TX
|
30
|
84
|
94
|
163
|
Detroit, MI
|
81
|
94
|
130
|
131
|
Buffalo-Niagara Falls, NY
|
57
|
70
|
70
|
82
|
Port Huron, MI
|
30
|
60
|
68
|
81
|
El Paso, TX
|
20
|
39
|
43
|
65
|
Otay Mesa, CA
|
0
|
19
|
24
|
35
|
Hidalgo, TX
|
5
|
13
|
18
|
26
|
Pembina, North Dakota
|
6
|
10
|
12
|
25
|
Nogales, AZ
|
7
|
14
|
14
|
23
|
Chicago, IL
|
2
|
4
|
7
|
23
|
|
U.S. trade with Mexico
|
92
|
203
|
233
|
391
|
10 top POE
|
76
|
199
|
227
|
383
|
Laredo, TX
|
30
|
84
|
94
|
163
|
El Paso, TX
|
20
|
39
|
43
|
65
|
Otay Mesa, CA
|
0
|
19
|
24
|
35
|
Hidalgo, TX
|
6
|
13
|
18
|
26
|
Nogales, AZ
|
7
|
14
|
14
|
23
|
Eagle Pass, TX
|
5
|
7
|
8
|
22
|
Santa Teresa, NM
|
0
|
1
|
1
|
18
|
Brownsville, TX
|
7
|
12
|
11
|
14
|
Calexico-East, CA
|
0
|
8
|
11
|
13
|
Del Rio, TX
|
2
|
2
|
3
|
4
|
Source: Authors’ own with data from the U.S. Department of Transportation Research and Innovative Technology Administration, Bureau of Transportation Statistics.
17Between 1995 and 2012, truck crossings between Mexico and the U.S. have grown by more than 78%. Throughout that period Laredo has maintained itself as the principal port of entry between the two countries and increased its share of total crossings from 26% in 1995 to 35% in 2012. Figure 2 reports the total number of truck crossings between the U.S. and Mexico, as well as the number of truck crossings for Laredo. In 2012, Laredo established itself as the number one NAFTA port of entry with 1.8 million truck crossings, ahead of Detroit, MI, which registered 1.5 million truck crossings.
18In contrast, railway crossings between the U.S. and Mexico have stagnated over the period 1995-2012 and Laredo’s share of that traffic has gone down, from 52% in 1995 to 39% in 2012. Figure 3 reports the total number of railway crossings between the U.S. and Mexico, as well as the number of railway crossings for Laredo. Despite its apparent decline, Laredo remains the principal port of entry for railway crossings between the U.S. and Mexico. With 3,492 railway crossings in 2012, Laredo was in fact the second most important in the NAFTA region, just behind International Falls, MI, with 3,924 crossings. Table 7 presents our complete dataset on truck and railway crossings for the NAFTA region.
19Laredo’s preponderance as a NAFTA POE in terms of both the value of trade as well as the physical magnitude of truck and railway crossings points to yet another notable regional feature of U.S.-Mexico trade: the continental relevance of the Texas-Northeastern Mexico region.
Figure 2. Truck crossings, Laredo, TX and U.S.-Mexico total, 1995, 2000, 2005 and 2012
Source: Authors’ own with data from the U.S. Department of Transportation Research and Innovative Technology Administration, Bureau of Transportation Statistics.
Figure 3. Railway crossings, Laredo, TX and U.S.-Mexico (total), 1995, 2000, 2005 and 2012
Source: Authors’ own with data from the U.S. Department of Transportation Research and Innovative Technology Administration, Bureau of Transportation Statistics.
Table 7. Truck and Railway Border Crossings, U.S.-NAFTA and U.S.-Mexico Trade, 1995, 2000, 2005 and 2012
1995
|
Rank
|
Truck
|
Railway
|
Port of entry
|
Crossings
|
Port of entry
|
Crossings
|
|
Top 5 ports: U.S.-NAFTA
|
4,127,621
|
Top 5 ports: U.S.-NAFTA
|
18,722
|
1
|
Detroit, MI
|
1,206,477
|
Laredo, TX
|
4,925
|
2
|
Buffalo-Niagara Falls, NY
|
948,682
|
Port Huron, MI
|
3,744
|
3
|
Laredo, TX
|
747,241
|
International Falls, MN
|
3,431
|
4
|
Port Huron, MI
|
618,479
|
Detroit, MI
|
3,368
|
5
|
El Paso, TX
|
606,742
|
Buffalo-Niagara Falls, NY
|
3,254
|
|
Total crossings U.S.-NAFTA
|
7,995,635
|
Total crossings U.S.-NAFTA
|
40,453
|
|
Top 5 ports, % of total
|
51.6
|
Top 5 ports, % of total
|
46.3
|
|
|
|
|
|
|
Top 10 ports: U.S.-Mexico
|
2,721,778
|
Top 9 ports: U.S.-Mexico
|
9,432
|
1
|
Laredo, TX
|
747,241
|
Laredo, TX
|
4,925
|
2
|
El Paso, TX
|
606,742
|
El Paso, TX
|
1,431
|
3
|
Otay Mesa/San Ysidro, CA
|
445,770
|
Eagle Pass, TX
|
1180
|
4
|
Brownsville, TX
|
223,689
|
Brownsville, TX
|
631
|
5
|
Nogales, AZ
|
206,032
|
Nogales, AZ
|
456
|
6
|
Hidalgo, TX
|
177,459
|
Otay Mesa/San Ysidro, CA
|
436
|
7
|
Calexico East, CA
|
175,983
|
Calexico East, CA
|
264
|
8
|
Eagle Pass, TX
|
53,026
|
Presidio, TX
|
101
|
9
|
San Luis, AZ
|
44,455
|
Tecate, CA
|
8
|
10
|
Tecate, CA
|
41,381
|
|
|
|
Total crossings U.S.-Mexico
|
2,860,625
|
Total crossings U.S.-Mexico
|
9,432
|
|
Top 10 ports, % of total
|
95.1
|
Top 9 ports, % of total
|
100.0
|
2000
|
Rank
|
Truck
|
Railway
|
Port of entry
|
Crossings
|
Port of entry
|
Crossings
|
|
Top 5 ports: U.S.-NAFTA
|
6,020,153
|
Top 5 ports: U.S.-NAFTA
|
19,185
|
1
|
Detroit, MI
|
1,769,389
|
Port Huron, MI
|
5,406
|
2
|
Laredo, TX
|
1,493,073
|
Detroit, MI
|
3,919
|
3
|
Buffalo-Niagara Falls, NY
|
1,198,085
|
Buffalo-Niagara Falls, NY
|
3,704
|
4
|
Port Huron, MI
|
839,200
|
International Falls, MN
|
3,456
|
5
|
El Paso, TX
|
720,406
|
Laredo, TX
|
2,700
|
|
Total crossings U.S.-NAFTA
|
11,573,707
|
Total crossings U.S.-NAFTA
|
40,555
|
|
Top 5 ports, % of total
|
52.0
|
Top 5 ports, % of total
|
47.3
|
|
|
|
|
|
|
Top 10 ports: U.S.-Mexico
|
4,339,710
|
Top 8 ports: U.S.-Mexico
|
7,108
|
1
|
Laredo, TX
|
1,493,073
|
Laredo, TX
|
2,700
|
2
|
El Paso, TX
|
720,406
|
Eagle Pass, TX
|
1,448
|
3
|
Otay Mesa/San Ysidro, CA
|
688,340
|
El Paso, TX
|
970
|
4
|
Hidalgo, TX
|
374,150
|
Nogales, AZ
|
774
|
5
|
Brownsville, TX
|
299,238
|
Brownsville, TX
|
694
|
6
|
Calexico East, CA
|
278,811
|
Calexico East, CA
|
241
|
7
|
Nogales, AZ
|
254,694
|
Otay Mesa/San Ysidro, CA
|
204
|
8
|
Eagle Pass, TX
|
106,892
|
Tecate, CA
|
77
|
9
|
Tecate, CA
|
62,878
|
|
|
10
|
Del Rio, TX
|
61,228
|
|
|
|
Total crossings U.S.-Mexico
|
4,525,579
|
Total crossings U.S.-Mexico
|
7,108
|
|
Top 10 ports, % of total
|
95.9
|
Top 8 ports, % of total
|
100.0
|
2005
|
Rank
|
Truck
|
Railway
|
Port of entry
|
Crossings
|
Port of entry
|
Crossings
|
|
Top 5 ports: U.S.-NAFTA
|
6,006,391
|
Top 5 ports: U.S.-NAFTA
|
20,303
|
1
|
Detroit, MI
|
1,745,318
|
Port Huron, MI
|
6,344
|
2
|
Laredo, TX
|
1,455,607
|
International Falls, MN
|
3,980
|
3
|
Buffalo-Niagara Falls, NY
|
1,142,411
|
Detroit, MI
|
3,602
|
4
|
Port Huron, MI
|
922,401
|
Laredo, TX
|
3,459
|
5
|
El Paso, TX
|
740,654
|
Buffalo-Niagara Falls, NY
|
2,918
|
|
Total crossings U.S.-NAFTA
|
11,459,841
|
Total crossings U.S.-NAFTA
|
42,265
|
|
Top 5 ports, % of total
|
52.4
|
Top 5 ports, % of total
|
48.0
|
|
|
|
|
|
|
Top 10 ports: U.S.-Mexico
|
4,470,066
|
Top 10 ports: U.S.-Mexico
|
9,458
|
1
|
Laredo, TX
|
1,455,607
|
Laredo, TX
|
3,459
|
2
|
Otay Mesa/San Ysidro, CA
|
740,654
|
Eagle Pass, TX
|
1,812
|
3
|
El Paso, TX
|
730,253
|
El Paso, TX
|
1618
|
4
|
Hidalgo, TX
|
491,077
|
Brownsville, TX
|
1045
|
5
|
Calexico East, CA
|
320,212
|
Nogales, AZ
|
785
|
6
|
Nogales, AZ
|
266,233
|
Calexico East, CA
|
415
|
7
|
Brownsville, TX
|
234,640
|
Otay Mesa/San Ysidro, CA
|
239
|
8
|
Eagle Pass, TX
|
97,729
|
Tecate, CA
|
73
|
9
|
Tecate, CA
|
69,586
|
Presidio, TX
|
12
|
10
|
Santa Teresa, NM
|
64,075
|
Hidalgo, TX
|
0
|
|
Total crossings U.S.-Mexico
|
4,675,897
|
Total crossings U.S.-Mexico
|
9,458
|
|
Top 10 ports, % of total
|
95.6
|
Top 10 ports, % of total
|
100.0
|
2012
|
Rank
|
Truck
|
Railway
|
Port of entry
|
Crossings
|
Port of entry
|
Crossings
|
|
Top 5 ports: U.S.-NAFTA
|
5,774,810
|
Top 5 ports: U.S.-NAFTA
|
16,535
|
1
|
Laredo, TX
|
1,789,546
|
International Falls, MN
|
3,924
|
2
|
Detroit, MI
|
1,541,150
|
Laredo, TX
|
3,492
|
3
|
Buffalo-Niagara Falls, NY
|
940,221
|
Port Huron, MI
|
3,269
|
4
|
Otay Mesa, CA
|
778,929
|
Detroit, MI
|
3,073
|
5
|
El Paso, TX
|
724,964
|
Warroad, MN
|
2,777
|
|
Total crossings U.S.-NAFTA
|
10,727,430
|
Total crossings U.S.-NAFTA
|
37,481
|
|
Top 5 ports, % of total
|
53.8
|
Top 5 ports, % of total
|
44.1
|
|
|
|
|
|
|
Top 10 ports: U.S.-Mexico
|
4,886,892
|
Top 10 ports: U.S.-Mexico
|
8,957
|
1
|
Laredo, TX
|
1,789,546
|
Laredo, TX
|
3,492
|
2
|
Otay Mesa/San Ysidro, CA
|
778,929
|
Eagle Pass, TX
|
2,349
|
3
|
El Paso, TX
|
724,964
|
El Paso, TX
|
1392
|
4
|
Hidalgo, TX
|
481,620
|
Nogales, AZ
|
657
|
5
|
Calexico East, CA
|
322,424
|
Brownsville, TX
|
600
|
6
|
Nogales, AZ
|
307,626
|
Calexico East, CA
|
252
|
7
|
Brownsville, TX
|
218,187
|
Otay Mesa/San Ysidro, CA
|
215
|
8
|
Eagle Pass, TX
|
117,375
|
Hidalgo, TX
|
0
|
9
|
Santa Teresa, NM
|
80,744
|
Santa Teresa, NM
|
0
|
10
|
Del Rio, TX
|
65,477
|
Del Rio, TX
|
0
|
|
Total crossings U.S.-Mexico
|
5,103,925
|
Total crossings U.S.-Mexico
|
8,957
|
|
Top 10 ports, % of total
|
95.7
|
Top 10 ports, % of total
|
100.0
|
Source: Authors’ own with data from the U.S. Department of Transportation Research and Innovative Technology Administration, Bureau of Transportation Statistics.
20As shown previously, Texas and the Northeastern States of Mexico (Chihuahua, Coahuila, Nuevo León and Tamaulipas) account for a significant share of U.S.-Mexico trade flows. This section focusses on economic integration in the Texas-Northeastern Mexico region and its relationship with trade. For this purpose, we present two distinct lines of evidence. The first consists of a statistical model that captures the historical dynamics of economic integration between Texas and Northeastern Mexico and allows us to test for the impact of NAFTA on that region. The second line of evidence comes from an original regional input–output model that quantifies the contribution of Texas to the Northeastern Mexico economy.
21Statistical correlations between macroeconomic quantities such as Gross Domestic Product (GDP), employment or wages have been widely employed to assess the level of economic integration between national or regional economies (Frankel and Ross 1998, Clark and Van Wincoop 2001, Hanson 1996, 2001, Yoskowitz et al. 2002, Ayala Gaytán and Chapa Cantú 2007, Garduño Rivera 2014).
22The business cycle synchronization model we present here measures correlations through time between GDP growth rates for Texas and Northeastern Mexico. The analysis is carried out with a Vector Auto-Regressive model (VAR) that also includes growth rates for the rest of Mexico as well as the United States. The model is specified as:
In (1), the “Δy” terms represent a GDP growth rate for a particular economy identified by the first sub-index: N= Northeastern Mexico, R=Rest of Mexico, U=United States and T=Texas, for a specific year “t”. Current growth rates for “N” and “R” (left-hand side of 1) are regressed on: a constant term, lagged values of growth rates for “N” and “R” (where “k” is the number of lags), the current growth rate for “U”, the current growth rate for “T”, and an additional term that combines the current growth rate in Texas with a dummy variable “Dht” which takes the value of 1 in year “t” if “t” is larger or equal to year “h”, and 0 otherwise. This latter term is designed to capture NAFTA’s impact on the relationship between growth rates; how “h” was empirically determined is explained later. Finally, the last term on the right hand side of (1) is a statistical error term. 23The VAR procedure applied to (1) produces numerical estimates and statistical significance measures for the coefficients “αij(k)”, “βi”, “δi” and “θi”, i=(1,2)=(N,R) and j=(1,2)=(N,R). The data used consists of official state annual GDP estimates from 1940 to 2011, supplemented where needed for Northeastern Mexico by previously published estimations (German Soto 2005, Ayala Gaytán et al. 2011).
24Preliminary work confirmed two necessary conditions required of the data in order to proceed with the VAR procedure. First, we verified that all four annual GDP growth rate series were stationary I(0); as usually is the case with this type of data, the original GDP data proved non-stationary I(1) and first-differencing was required to solve that problem. Second, we verified that Texas and United States GDP growth rates were exogenous (in the Granger sense) to Northeastern Mexico and Rest of Mexico growth rates; this justifies including the former variables in the right-hand side of (1).
25Next in order to determine the dummy year “h”, we ran our VAR model for all possible structural change dummy variables from year 1986 to 2000, and selected the year in which at least one θi coefficient was significantly different from zero and the sum of square of the residuals was the lowest possible. The result obtained was h=1997. We take this year (which interestingly coincides with the end of the 1995 Mexican financial crisis) as the one in which the effect of NAFTA kicked in. Now the direction of this effect and whether it is significant or not depends on the estimates for the θi coefficients obtained from the VAR procedure. For example, if θi>0 and proves statistically significant, we can conclude that for economy “i”, NAFTA increased the level of business cycle synchronization (i.e. economic integration).
26The last step before proceeding with the VAR estimation was to determine the number of lags to be included in (1). The standard Likelihood Ratio Test and the multivariate Akaike Information Criteria suggested k=4, however we later eliminated the third lag because none of the coefficients associated with this lag were significant.
27Table 8 reports the VAR coefficients estimates. These results reveal a significant level of business cycle synchronization between Texas and Northeastern Mexico, as well as between Texas and the Rest of Mexico: in both cases, the estimated “δi” are highly significant statistically. However, for the former pair the relationship is almost twice as strong as for the latter (0.90 vs 0.55). Moreover, the effect of NAFTA although significant in both cases (with statistically significant estimated “θi”) is stronger for Northeastern Mexico (0.65) than for the Rest of Mexico (0.43). These results suggest that more regional trade has led to heightened cross-border economic integration for Northeastern Mexico.
Table 8. VAR estimates of model (1) coefficients (standard errors in parenthesis) and regression statistics
|
ΔYN (t)
|
ΔYR (t)
|
Constant
|
-0.006783
(0.01345)
|
-0.007482
(0.00958)
|
ΔYN (t-1)
|
0.602141**
(0.29140)
|
0.161495
(0.20758)
|
ΔYN (t-2)
|
-0.685075**
(0.26685)
|
-0.466041**
(0.19010)
|
ΔYN (t-4)
|
-0.842740***
(0.27303)
|
-0.504075**
(0.19450)
|
ΔYR (t-1)
|
-1.012730**
(0.43554)
|
-0.232402
(0.31027)
|
ΔYR (t-2)
|
1.007991**
(0.40914)
|
0.738044**
(0.29146)
|
ΔYR (t-4)
|
0.785447*
(0.40780)
|
0.605549**
(0.29051)
|
ΔYU (t)
|
0.382366
(0.27754)
|
0.262953
(0.19771)
|
ΔYT (t)
|
0.902597***
(0.20516)
|
0.551532***
(0.14615)
|
ΔYT (t)*D(h=1997)
|
0.649328**
(0.30575)
|
0.430035*
(0.21781)
|
|
R2
|
0.627781
|
0.643747
|
Adjusted R2
|
0.539624
|
0.559372
|
Sum of Squared of Residuals
|
0.042883
|
0.021762
|
F statistics
|
7.121167***
|
7.629543***
|
Note: *** p<0.01, ** p <0.05 & * p<0.1. Source: authors’ own calculations.
28To quantify the economic contribution of Texas to the Northeastern Mexico economy, we begin with an original input-output (I-O) model for Northeastern Mexico, which builds on and expands previously published work (Ayala Gaytán et al. 2008, 2009a, 2009b). It includes the states of Coahuila, Nuevo León and Tamaulipas (Chihuahua therefore is left out) and is calibrated for year 2004.
29The I-O model includes 17 economic sectors of Northeastern Mexico. Additionally, for seven of those sectors the model distinguishes explicitly the portion that operates as maquiladora. This form of production combines local labor with exclusively foreign intermediate goods, for the sole purpose of exporting finished goods. In Table 9 these maquiladora sector portions are identified as “assembly”. For example, in Table 9 the “Metallic Products, Machinery and Equipment” entry, which includes the automotive, electrical and electronics industries, refers to the portion of that sector that relies at least to an extent on domestic intermediate goods and produces at least in part for the domestic market, while the “Metallic Products, Machinery and Equipment (assembly)” entry refers to the maquiladora portion of the sector.
30We added the Texan economy to the I-O model based on an analysis of commercial flows between Northeastern Mexico and Texas by economic sector, using the North American Transborder Freight Database (U.S. Bureau of Transportation Statistics). For that purpose, we treated commercial flows between Northeastern Mexico and Texas as endogenous and assumed that each of the Northeastern Mexico economic sectors requires Texas’ products as intermediate inputs (imports) and each of the Texas economic sectors buys Northeastern Mexico’s products (exports). The structure of the full I-0 model is as follows:
31In (2) “x” is a 25x1 vector that contains the Gross Output (GO) of each of the 24 Northeastern Mexico economic sector or sector portion as defined earlier; the last element is the value of the imports from Texas. The 25x1 vector “y” contains the final demands of each of the Northeastern economic sector after subtracting their exports to Texas. The 25x25 matrix “A” is the technical coefficient matrix that reports requirements of intermediate inputs for sector j from sector i; the last row includes the requirements of intermediate inputs for sector j from Texas, and the last column contains the distribution of the exports to Texas by economic sector (in proportional terms).
32Model (2) was fed the following scenario: a 100% decrease of imports from Texas in Northeastern Mexico and simultaneously, a 100% decrease of Northern Mexico’s exports to Texas. We use here the standard hypothetic extraction method (Wang et al. 2013, Llano 2009, Salgueiro et al. 2010, Song and Liu 2007). Comparing the gross output (GO) pre- and post-scenario for each of the model’s sectors allows us to quantify the contribution of Texas to the Northeastern Mexico economy.
33Table 9 presents the results of this exercise. Results reveal that 12.6% of Northeastern Mexico’s GO is related to trade with Texas, equivalent to USD 25,604 million. All of Northeastern Mexico’s economic sectors show some degree of linkage with the Texas economy; however the extent of that dependence is quite varied.
34The maquiladora portions of Northeastern Mexico’s economic sectors present a much higher than average level of linkages with Texas: between 26 and 28% of their GO depends on trade with Texas. Given the characteristics of these industries, this should come as no surprise. Nonetheless, the domestic portions of some of those same sectors show surprisingly strong linkages, especially in the case of manufacturing activities, for example Metallic Products, Machinery and Equipment (9.3%) and Textile (7%). As should be expected, linkages are quite low for all four domestic service sectors (Commerce, restaurant and hotels; Communications and transport; Financial and real estate services; Social and personal services), with considerably less than 1% of their GO linked to trade with Texas.
35The results from our two models complement each other and shed additional light on the nature of the border effect. For the Texas-Northeastern Mexico region, the border effect has meant much more than grabbing a relatively large share of overall U.S.-Mexico trade. It has led to a significant degree of economic integration in a macroeconomic, business cycle synchronization sense. However, as our I-O analysis reveals, this integration process has not been across the board. On the contrary, the border effect has selectively and disproportionally played itself out within specific sectors and types of economic activities.
Table 9. Contribution of Texas to the Northeastern Mexico economy, by sector, 2004 (millions of constant pesos)
Economic sector
|
Initial GO
|
GO without Texas
|
Effect (%)
|
Agriculture, livestock, forestry and fishing
|
39,990
|
38,960
|
-2.6%
|
Mining
|
26,467
|
21,118
|
-20.2%
|
Food, beverages and tobacco
|
114,804
|
113,512
|
-1.1%
|
Textile industry
|
14,188
|
13,189
|
-7.0%
|
Wood industry
|
4,440
|
4,330
|
-2.5%
|
Paper industry
|
15,615
|
15,253
|
-2.3%
|
Chemical industry
|
76,654
|
73,441
|
-4.2%
|
Non Metallic Mineral Products
|
51,900
|
50,715
|
-2.3%
|
Basic Metallic industries
|
85,212
|
80,089
|
-6.0%
|
Metallic Products, Machinery and Equipment
|
363,390
|
329,501
|
-9.3%
|
Other manufacturing industries
|
5,287
|
4,712
|
-10.9%
|
Building
|
87,061
|
87,061
|
0.0%
|
Electricity, gas and water
|
51,382
|
50,541
|
-1.6%
|
Commerce, restaurant and hotels
|
284,523
|
282,950
|
-0.6%
|
Communications and transport
|
157,558
|
156,873
|
-0.4%
|
Financial and real estate services
|
134,318
|
133,869
|
-0.3%
|
Social and personal services
|
303,781
|
302,702
|
-0.4%
|
Food, beverages and tobacco (assembly)
|
1,387
|
1,026
|
-26.0%
|
Textile industry (assembly)
|
31,563
|
22,257
|
-29.5%
|
Wood industry (assembly)
|
4,248
|
3,114
|
-26.7%
|
Chemical industry (assembly)
|
10,178
|
7,423
|
-27.1%
|
Metallic products, machinery and equipment (assembly)
|
234,290
|
172,781
|
-26.3%
|
Other manufacturing industries (assembly)
|
9,572
|
6,899
|
-27.9%
|
Services (assembly)
|
5,248
|
3,874
|
-26.2%
|
Texas
|
148,617
|
0
|
-100.0%
|
Total
|
2,261,675
|
1,976,190
|
-12.6%
|
Source: authors’ own calculations.
36Trade between Mexico and the United States has increased substantially under NAFTA. The evidence presented here shows that the regional distribution of this trade has followed varied and multi-layered border and trans-border dynamics. On the one hand, the Border Region (in particular, the Texas-Northeastern Mexico region) has received a large share of the trading spurred by NAFTA. On the other hand, other regions of Mexico (in particular, Central Mexico) have also participated to a significant degree in the process as well as have some non-border U.S. states, such as Michigan. No doubt, geography does matter in shaping economic relations: the simple logic of geography binds nations together (Wilson and Biette 2014). Nevertheless other factors are also obviously at play. For example, culture certainly has played a role in the historical development of economic relations between Texas and Northeastern Mexico (Aguilar Barajas 1995). What remains clear from our results is that the way trade flows between the U.S. and Mexico have evolved at a regional level reflects much more than solely a border effect.
37Our results further suggest that NAFTA has to a good extent reinforced pre-Agreement patterns. In the case of the Texas-Northeastern Mexico region, before NAFTA, Northeastern Mexico’s economy was already more integrated with Texas than was the case of the Rest of Mexico; NAFTA only increased this differential. Economic integration and economic linkages in the Texas-Northeastern Mexico region also illustrate another aspect of the way the border effect has played out in this case. Underneath the relatively high level of business cycle synchronization displayed in the region, lurks a very heterogeneous pattern of cross-border relationships between economic sectors and types of economic activities.
38Important policy implications derive from these findings. Increasing trade has brought about the need for better ways to manage the common border region. Strains on border crossing facilities and needed improvements in infrastructure and customs formalities have been long identified as challenges for the 21st Century (TRB 2003). Hence trade corridors and infrastructure in Mexico are central to the enhancement of trade between the two countries. As found almost a decade ago (Casares and Sobarzo 2004), improvements in transport infrastructure in Mexico correlate positively with higher volumes of trade with the United States. Our findings suggest that such improvements should be carefully targeted. In particular, improvements in the functioning of the Laredo Port Of Entry, which is by far the major POE between the U.S. and Mexico, would generate wide benefits on both sides of the border.
39Our findings support several recommendations expressed in the context of NAFTA’s 20th anniversary. One of those recommendations has to do with the need to increase regional competitiveness through economic cooperation and integration (Wilson and Biette 2014). In light of what has been discussed, the challenges of integration also mean border and implicitly urban management challenges.
40The unit of analysis in this paper is the state. However at the end of the day, trade happens between specific, generally urban locations and not wide geographical areas, hence the importance of metropolitan areas and major urban centers in the U.S.-Mexico border region. There is the implicit assumption that state-level trade patterns reflect the pattern of trade at the urban level. The case of Texas and the Northeastern Mexico illustrates the point. It is not Chihuahua as a state but rather Ciudad Juárez, a major border export-orientated assembly hub, as well as the Chihuahua state capital, that have links with Texan cities such as Dallas, Houston, San Antonio and Austin. The same applies for Nuevo León, where industry is almost entirely concentrated within the Monterrey Metropolitan Area, which is at the center of trade with Texas. Recent work has shown that Northern Mexico urban areas close to the U.S. border, especially the largest ones, benefitted more from NAFTA than cities located further away from the border (Garduño Rivera 2014).
41Clearly there remains much to be studied about regional trade patterns within North America at different geographical scales. As knowledge increases about the characteristics of these patterns, their historical evolution and consequences in terms of regional inequalities, development and economic integration, the better for the design of effective and fair regional development policies.