The Origins and Progress of U.S.-Mexican Trade, 1825–1884: “Hoc opus, hic labor est”

The economic and commercial history of early national Mexico remains very much a mystery, the contributions of John Coatsworth, Donald Stevens, Barbara Tenenbaum, Guy Thomson, David Walker, and numerous Mexican scholars notwithstanding.¹ In part, the problem is one of finding, collating, and interpreting statistics that are frequently difficult, conjectural, or contradictory. But it is also a question of knowing where to start; of deciding what is important; and of creating a framework within which statistics and other economic data assume coherence, consistency, and meaning, even in the presence of incomplete information.

A likely point of departure is foreign trade. It is clear from the work of Stevens, Tenenbaum, and Thomson that foreign trade is central to discussions of political change, capital movements, indebtedness, and commercial policy. Foreign trade was nearly at the heart of early national political economy. But the composition, evolution, and effects of foreign trade are by no means well known, much less undisputed. Tenenbaum links the fragility of the fiscal system, and indirectly of federalism itself, almost entirely to the cyclical variability of trade and tariffs. Stevens, on the other hand, discerns a politicized economic cycle and reverses cause and effect, concluding that “politicians [in Mexico] did not merely respond to economic cycles, but caused them.”² Clearly then, examining foreign trade is a means of understanding issues that largely defined the existence, character, and viability of the Mexican state in the early national period. “Sin hacienda, no hay estado,” as a publication of the day remarked.³

Yet a study of Mexico’s foreign trade in the early nineteenth century is also necessarily an analysis of its commercial policy. And commercial policy was, and is, a weapon. It was, perhaps, uniquely effective against the pressures that both Great Britain and the United States exerted on Mexico, for the Mexican market was an object of competition between them, and competition brings leverage. Mexico employed the weapon, sometimes successfully and sometimes less so, but always in reaction to enormous pressures on its sovereignty. In Mexican eyes, the flag followed trade.

From a historical standpoint, then, studying early national trade patterns and commercial policy allows us to draw large lessons about the behavior of small polities and to discuss a neglected aspect of Mexican history as well. But where to begin? The Anglo-Mexican trade was of paramount importance, but before 1858 Great Britain ostensibly kept no systematic account of its bullion imports.⁴ Pending further research, Britain’s early balance of trade with Mexico is mostly a matter of conjecture. The Franco-Mexican trade was not unimportant, but before 1847 its statistics reflect “official values.”⁵ So we begin with the record of United States–Mexican trade.

While the record of the United States–Mexican trade is far from perfect, it is nevertheless voluminous and can be adjusted for the errors and omissions characteristic of all trade data. Moreover, the trade cycle with the United States is at once typical and peculiar, and the tension itself is revealing. Trade deficits, one may say, are trade deficits. They vary in size but differ in degree rather than in kind. In this sense, trade with the United States was no less representative than any other. Yet this trade also carried unique implications for both nations and was subject to social, political, and diplomatic pressures that the Anglo- and Franco-Mexican trades never faced. In other words, “ordinary” exchange between the United States and Mexico sometimes reflected unusual circumstances and at times produced untoward results.

Finally, the size of trade between Mexico and the United States may well have been “small” in a conventional sense, but its impact depends less on its absolute size than on the scale of measurement chosen.⁶ In the Mexican view, any trade with the United States could be potentially hazardous, whatever its volume. From the standpoint of the United States, the silver that Mexico used to finance its purchases had direct effects on the U.S. money supply into the 1830s, and indirectly thereafter. Mexican silver fueled the inflation of the middle 1830s in the United States and was a major cause of the Panic of 1837.⁷ Moreover, the trade to Mexico had important regional consequences within the United States. As late as 1830, Philadelphia merchants believed that they handled about 50 percent of the U.S. trade to Mexico, and Mexico was the city’s fifth largest trading partner overall.⁸ How much the trade counted is a matter of perspective.

Exports, Imports, and the Balance of Trade

To tell a coherent story, we need to know the U.S. balance of visible trade with Mexico, that is, the relative size of merchandise exports and imports. But the numbers do not come easily. The existing trade statistics are usually misleading and often incorrect. And the United States controlled its own carrying trade to Mexico, which means that its earnings from these services need to be taken into account, even though they are generally ignored.⁹ The ostensible balance of Mexico’s visible trade with the United States is probably wrong. The current account balance, that is, net income from trade, investment, and services, is a mystery.

The reliability of trade statistics is an enduring problem for historians, analysts, and policy makers; no discussion based on quantitative evidence can ignore these issues.¹⁰ Since trade is fundamentally a quantitative matter, getting the numbers right is a necessary first step. Thus, the discussion of sources and methods that follows is a central aspect of this essay, but it is a difficult one as well. “Hoc opus, hic labor est,” as a U.S. consul of the time complained when asked to provide similar commercial data.¹¹ For the sake of readability, I have relegated purely technical matters to four appendixes. Anyone wishing to replicate or appraise my results should find the data in the appendixes helpful.

The standard U.S. sources for imports from and exports to Mexico are Series U321 (exports) and U339 (imports) of the Historical Statistics of the United States.¹² Series U321 gives the current value of exports, noting only that “reexports” (total imports less retained imports) are included. This is an enormous understatement. Before 1841, in no year did reexports to Mexico comprise less than half of all U.S. exports by value. Even after 1841, reexports were hardly insignificant. In both 1852 and 1860 and even as late as 1872—when they were swelled by English materials for Mexican railroad construction—reexports were again 40 percent of all exports by value.¹³ Total and domestic exports to Mexico were by no means identical, as shown in Table 1.

The importance of reexports from the United States was well known. As late as 1852, the U.S. consul in Veracruz could remark that “the cargoes of our New York packets [consist] almost wholly of bonded goods from Europe and China.”¹⁴ Mexican geography and the early concentration of foreign commercial houses in Veracruz enhanced the original significance of the reexports trade. For example, the British commercial houses of Marshall & Manning [later Manning & Mackintosh]; Bates Barton & Co.; and Exter, Greaves & Co. [McCalmont, Greaves & Co.] were established in Mexico City and Veracruz.¹⁵ But the U.S. consul at Veracruz noted that “English goods going into Mexico through Texas . . . will injure the English trade [through Veracruz] exceedingly for the United States . . . can get them into the interior much cheaper than they can be transported from [Veracruz] up these everlasting hills.”¹⁶ Overland reexports through the United States also reflected the continuing productivity of Mexico’s northern silver mines in the wake of insurgent damage to the Guanajuato district in the 1810s. Before 1846, northern silver also drove the Santa Fe trade and financed substantial U.S. reexports to Chihuahua and Durango. To the extent that overland reexports through the United States reduced transportation costs, or to the extent that maritime reexports bypassed Veracruz for Alvarado or Tampico, the trade that had customarily linked Veracruz to Mexico City was diminished. As the English diplomat H. G. Ward observed, the British were committing the very error that the Spaniards had made in concentrating their energies on Veracruz and Mexico City. The immediate beneficiaries of this strategy were the U.S. merchants who acted as commercial intermediaries. Geography was thus the unspoken ally of the Yankee trader.¹⁷

For simplicity, we define “U.S. exports” as domestic exports (i.e., produced in the United States). We assume that reexports from the United States remained foreign property, but that U.S. merchants and shippers profited from carrying them to Mexico. In other words, U.S. reexports were largely British and French exports transported by U.S. carriers. They produced “invisible earnings” for the U.S. current account¹⁸ but did not affect the balance of its visible (merchandise) trade with Mexico.

In Table 2, column A, I give domestic exports to Mexico from 1824/. 25 to 1883/84. Two figures are given for 1862/63 through 1864/65. The first figure is “official” and recorded in American Commerce. Commerce of South America, Central America, Mexico, and West Indies, With Share of the United States and Other Leading Nations Therein, 1821–1898 (Washington, 1899), a source for Series U321 (exports to Mexico) and U339 (imports from Mexico) in the Historical Statistics of the United States. The bracketed second figure is corrected for contraband war materiel that flowed from Union ports to the Confederate States through Matamoros, Mexico during the U.S. Civil War. (See  Appendix A for details.) Exports were valued on a “free alongside” (FAS) basis (before adding insurance, freight, and merchants’ commissions). Nevertheless, all these numbers (and those that follow) are approximations and cannot be considered exact. The U.S. Constitution (section 9, clause 5) prohibits federal imposition of an export tax, so the government had no vested interest in consistently and correctly recording the value of exports. And well into the 1850s, the U.S. consul at Veracruz downplayed the precision of the statistics of the trade.¹⁹ But as we will see, this concern need not be exaggerated. Internal evidence suggests that the export totals are broadly correct. However we revise, transform, or manipulate the historical statistics, we uncover no substantial discrepancy between trends discussed in the consular reports and those indicated by the historical statistics. This strongly suggests that the U.S. consuls, whatever their personal financial interests, were reasonably unbiased reporters of economic information, or that their biases and interests did not significantly affect the reliability of the information they reported.²⁰

In column A (Table 2) I give exports in current prices, and their movements include changes in both volume and value. But we must distinguish between changes in export prices and export quantities. The usual way of doing so is by “deflating” a series in current prices, thus converting “nominal” into “real” values. For this conversion, we require an index of prices of U.S. exports to Mexico, or a good approximation of such an index. The Warren-Pearson index of U.S. wholesale prices spans the nineteenth century, but there is no guarantee that it accurately reflects the composition of exports to Mexico. The safest course is to construct our own export price index.

As I show in Table 3, the value share of the top five domestic goods exported to Mexico from the 1820s through the 1880s was always high— with the exception of the 1870s, it was never less than 50 percent—and in the 1830s and 1840s it was well over 60 percent. These goods included finished cottons, wheat flour, raw cotton, and after 1868/69 steel rails as a proxy for manufactures of iron and steel. Concentration makes the construction of the index much easier. I provide details of how the index was constructed in  Appendix B. Here we need only say that the deflator is a Laspeyres index with a base period of 1840/41–1844/45. Its weights were adjusted (and linked) in 1840/41, 1844/45, 1859/60, and 1868/69 to reflect the changing composition of exports. The index covers roughly 40 to 70 percent of all U.S. exports to Mexico by value, which is adequate. In Table 2, the index appears in column B. Real exports in constant 1840/41–1844/45 prices are given in column C and are nothing more than (A/ B × 100).

U.S. imports from Mexico are easier to measure. By and large, Mexico shipped precious metals—mostly silver—to the United States. Most came in specie, but there was some bullion as well. As I show in Table 4, specie and bullion (including some gold) were never less than half of all imports by value. Until its world price fell in the later 1870s and early 1880s, silver was often 60 to 70 percent of the value of all U.S. imports from Mexico. Without much error, then, we can deflate these imports (silver plus all others) by an index of silver prices constructed on an 1840/41–1844/45 base.²¹ Imports in current prices appear in column D, Table 2, and the silver price index in column E. Real imports in prices of 1840/41–1844/45 appear in column F and are nothing more than (D/E × 100).

Mexico’s export figures for precious metals are highly suspect, since substantial Mexican taxes on the coinage and export of specie during much of the nineteenth century made smuggling silver out of the country big business.²² But the U.S. import totals are probably accurate, even though silver and gold were admitted duty free; the absence of taxation made concealing them or evading customs pointless. While U.S. import accounts had to be reformed in the 1840s because of inaccuracies, their precious metal totals were considered basically sound.²³

With new series for United States imports from and exports to Mexico, we can now compute a more accurate balance of visible trade (merchandise exports less imports). In Table 2, column G, the balance appears in current dollars. It is simply net income from merchandise trade.

Not all international transactions are “visible.” Conventional balance-of-payments accounting distinguishes “visible trade” from “invisibles” such as trade in business services.²⁴ In general, the “balance on visible trade” differs from the “balance on current account” by the movement of “invisibles”. Practically speaking, the only computation of invisibles possible here is income from insurance, shipping, and commissions. I describe the procedure for calculating these in  Appendix C and credit to the United States the amounts computed as “invisible exports” to Mexico. The annual earnings from invisibles appear in Table 2, column H, and the balance on current account in column I, the sum of G plus H. The current account is simply net income from trade, services, and investment.

From Numbers to Notions: Silver, Cycles, and Tariffs

Numbers, of course, are not the whole story, but without them there would be no story. And the story they tell is one of both persistence and change. Mexico continued to trade silver for cloth, much as the Indies had under the Bourbon monarchs but with new wrinkles. As wheat fields in the United States began to appear west of the Alleghenies, New Orleans could supply Mexico with wheat flour shipped cheaply down the Mississippi. And then there was raw cotton. In the early 1840s, Mexico’s new cotton industry drew heavily on raw cotton imported from the United States. As Lucas Alamán observed in 1846, “Without [it], the factories could hardly have made half of what they did during the past two years.” The available evidence suggests that Alamán’s calculations were very nearly correct.²⁵

Yet there is a sense in which what did not change was as impressive as what did. The Mexican cotton industry in the 1830s and 1840s was based less on comparative advantage than on restrictions on international trade, a nostrum peddled by Mexican industrialists who found domestic markets difficult to control. From the standpoint of international trade cycles, moreover, the Mexican staple remained silver, and the impact of its production and export during the early and middle decades of the nineteenth century was substantial. There are two ways of seeing this.

First, and most obviously, silver was a medium of international exchange and figured prominently in Mexico’s capacity to import. The value of Mexican imports from the United States conformed closely to the value of the silver exported to the United States. Indeed, the correspondence is nearly exact.²⁶

Second, Mexico was a nation with small, fragmented capital markets that rationed credit through kinship networks rather than through banks or other formal intermediaries. Under these circumstances, development economists suggest, activities that raise investible wealth will have an economic impact disproportionate to their size.²⁷ In Mexico, silver mining was one such activity. Compared to agriculture, mining’s share in national income was not large, but it proved a concentrated and unrivaled source of liquid, mobile savings to investors. As a result, and in the long run, variations in silver production were tied to variations in the supply of loanable funds. Thus Guy Thomson’s calculations of the volume of loans in Puebla between 1800 and 1830 show a decline whose severity could only be explained by an equally sharp fall in the production of silver.²⁸

As loans rise and fall, the volume of “real” economic activity financed by credit must also grow and contract, and so too will imports and exports. Thus, in an open economy, the balance of trade must reflect changes in the business cycle, and in Mexico the business cycle was necessarily tied to the credit that silver mining supplied. In early national Mexico, trade cycles, business cycles, and the production of silver had to be related.²⁹

The evidence from Mexico’s trade with the United States is consistent with these hypotheses. In logarithms, real U.S. imports from and exports to Mexico (see Figure 1) conform closely to each other, indicating comparable rates of change. Indeed, on the face of things, U.S. imports of silver, hides, dyewood, and indigo from Mexico move somewhat in advance of U.S. exports to Mexico, especially before the late 1860s. This pattern may suggest an “export-led” model of Mexican growth that persisted until the onset of the depreciation of silver after 1867 and until the beginnings of industrialization under Porfirio Díaz somewhat later.³⁰

Figure 1 reflects roughly three similar cycles in imports and exports. Measured from trough to trough, the first spans 1825/26–1847/48, with a peak in 1835/36. The second covers 1847/48–1867/68, with a peak around 1860/61. The last begins in 1867/68, peaks around 1872/73, and closes in 1883/84. The dangers of teleology notwithstanding, the chronology conforms to not a few significant points of Mexican history between Iturbide and Díaz: the Texas rebellion; the War of 1847; the French invasion (and U.S. Civil War); the Restored Republic; the death of Juárez and the accession of Lerdo de Tejada; and the completion of rail links with the United States. Of course, any year might well be invested with significance in those turbulent times. Yet more than coincidence is at work, for Mexican trade cycles were thoroughly politicized as well. Over the long run, cyclical changes in the production and export of silver clearly mattered. But in the shorter run, the impress of political factors is obvious.

Politics, Edward Nell observes, is economics pursued by other means.³¹ Since political factors were deeply embedded in the short-run movements of Mexican trade and commerce, who could disagree? But in this context “politics” meant two things: tariff policy, and the uncertainty that coups, blockades, pronunciamientos, wars, and repeated changes of regime engendered.

The level of tariff protection was clearly a political outcome. The private demand for industrial protection, the public demand for revenue, and the diplomatic aspects of trade and territorial expansion all played a role in setting the tariff. Moreover, the enforcement of tariffs was a free-for-all. If several tariffs were in force, which was to be honored? Did national, state, or local tariffs take precedence? Who, if anyone, knew the code or could even find a copy of it on demand? It was a nightmare for merchants and consuls, but chaos has its reasons, and uncertainty could be a potent ally for unstable regimes.

Economic uncertainty is an amorphous idea, akin, perhaps, to risk of indeterminate probability. But risk and uncertainty were key elements in the economic calculus of early national Mexico. Barracks revolts, pronouncing generals, extreme factionalism, blockades, and the risk of war paralyzed trade and commerce as effectively as did garbled regulations and unenforceable property rights. Merchandise no longer moved, debts went uncollected, and silver shipments slowed.

The First Cycle (1825/26-1847/48)

Texas dominated the political economy of the first cycle. Its trauma diverted Mexican trade toward Great Britain, whose role as a potential counterweight to U.S. territorial designs on Mexico ended only with annexation and the Mexican War. The theme appears in modern Mexican scholarship and pervades the writings of men as different as Alamán and Carlos María de Bustamante. The message is clear: the flag follows trade. Trading with the United States brings their merchants, and their merchants bring trouble. “They are the true sons of Englishmen,” wrote Bustamante, “whose example in India they remember and emulate. Merchants financed the invading expedition. Once their company had gotten hold of the land, they turned it over to the crown, which installed a government and set millions of Indians groaning under a slavery enforced by bayonets.”³² To judge the dramatic effect that the Texas rebellion had on the pattern of trade between Mexico, the United States, and Great Britain, see  Appendix D.

Alamán’s point was much the same. “Instead of armies, battles, and invasions, which raise such uproar and generally prove abortive, [the United States] use means which, considered separately, seem slow, ineffectual, and sometimes palpably absurd, but which united, and in the course of time, are certain and irresistible.” And what were these means? The list was a long one, but the first Alamán mentioned was “commercial negotiations.”³³

To be sure, Mexico had never trusted U.S. commercial ambitions. In 1829, Secretary of State Martin van Buren termed Mexico’s dilatory consideration of a treaty of reciprocity a “mistaken policy . . . unfriendly to the commercial prospect of the United States.” Mexico regarded the United States “with a degree of indifference and suspicion as extraordinary as it was to be regretted.” Negotiation over the treaty, which had begun in 1826, dragged on until 1832. And even then, “the first attempts of our adventurous citizens [were] burdened by the imposition of prohibitive duties . . . [in] Mexican ports.”³⁴

The tariff, then, was Mexico’s weapon of choice. Before the War of 1847, Mexico repeatedly adjusted its coverage and level, most notably in 1829, 1837, and 1842-43. In theory, the tariffs covered a variety of articles, but in practice their target was finished cottons, the industry that Mexican industrialists sought most strongly to protect. Things started badly in 1825, with allegations of discriminatory duties on U.S. cottons, and by 1827 rising duties had driven U.S. exports sharply down.³⁵ Matters worsened in 1829, and U.S. merchants in Veracruz warned that the new duty’s “pernicious influence . . . has annihilated the reviving spirit of commercial enterprise.” The U.S. consul concurred. When pressed, Lucas Alamán, then secretary of home and foreign relations, promised relief. But his suggesting that repeal of the duty on coarse cottons would pass the Mexican Senate “without the slightest opposition” was simply bad faith on Alamán’s part.³⁶

Much worse was in store. In the continuing wake of the Texas crisis, the tariff was again revised. Although some duties were reduced in 1837, the new schedule prohibited (effective March 1838) ordinary cottons and woolens, cotton yarn and thread, and ready-to-wear clothing. In late 1842, duties on goods otherwise permitted rose 30 to 50 percent. The Tariff of 1843 reiterated the prohibitions of 1837 but added raw cotton and coarse woolens to a list that included at least sixty articles “embracing most of the necessaries of life and far the greater portion of [U.S.] products and fabrics.”³⁷ As the U.S. consul in Veracruz remarked, “no cotton goods can be imported less than 25 and 30 threads [per quarter square inch] which comprises that very kind of good suited and worn by the poor and middling classes of the community.” Here was a recipe for reviving the moribund obrajes of the colonial regime.³⁸

The prohibitions were murderously effective. Before 1838, finished cottons were 30 to 40 percent of domestic U.S. exports. Once the Tariffs of 1837 and 1842 had taken hold, the share of cottons fell to only 16 percent (see Table 5). A small market for cotton twist, yarn, and thread was annihilated as well. “[Mexico’s] commerce would be infinitely important to us,” said the U.S. minister in 1842, “but for this unfortunate Texan war, which has caused much injury to the United States.”³⁹

In late 1845, a bill pending in the Chamber of Deputies would have admitted cotton and cotton manufactures on better terms. Seven percent of customs duties would indemnify cotton growers and manufacturers for their losses to foreign competition.⁴⁰ The bill failed; the tariff scheduled to go into effect in February 1846 was as restrictive as its predecessors. But the war intervened, and Mexican ports were placed under blockade.

The U.S. consul in Veracruz was no doubt correct when he observed in late 1845 that “Mexico never since she has been a Nation has been in so wretched a State.”⁴¹ But wretched is not powerless. Although an amalgam of economic nationalism and opportunism, Mexican policy nevertheless rested on the law of demand. This was the nation’s best weapon.

In the long run, Mexico’s strategy did not and probably could not have prevented the loss of Texas, New Mexico, and California to the United States. But in the short run it was hardly irrational. High tariffs satisfied the demand for protection that manufacturers pressed so insistently on the Mexican Congress and mollified other vital (and volatile) constituencies as well.⁴² And, indeed, trade with the United States remained small, much to the chagrin of those who had expected great things of the Mexican market. In the early days of the First Republic, newspapers in the United States hailed Guadalupe Victoria as another George Washington.⁴³ By 1845, the comparisons drawn were altogether less flattering.

A final observation. In the very short run, large fluctuations in trade occurred from year to year. Some were simply random, and not all are easily or equally explicable. Yet contemporaries understood the link between political stability and sustained growth. As one anonymous essayist put it: “The mere rumor of a revolution is pernicious.. . . Agriculture falls off, commerce is all but paralyzed, and silver shipments cease because the roads are probably not safe. In short, the citizens are in arms, and all is in disorder. These are the necessary and immediate consequences of the very rumor, more or less substantiated, of the next revolution.”⁴⁴ Uncertainty dominated yearly, and daily, affairs.

The Second Cycle (1847/48-1867/68)

Historians of the United States once called their Civil War the “irrepressible conflict.” No one familiar with relations between the United States and Mexico in the 1840s could conclude that the Mexican War was any less “irrepressible.” U.S. ambitions and Mexican nationalism were mutually exclusive. Indeed, the drive to commercial and territorial expansion characteristic of U.S. foreign policy in the 1840s has been termed “manifest design” by a historian who argues that this expansion was neither accidental nor providential.⁴⁵ In the long run, Mexico’s defeat (and the annexation of Texas) implied a permanent increase in the U.S. market and a sweeping reorientation of Mexico’s trade. I highlight this increase in Figure 2 by shifting the X-axis (years beginning in 1825/26) upward to intersect the log of real exports in 1847/48. Exports naturally rose and fell thereafter, but they rarely returned to antebellum levels. Commercial expansion may or may not have “caused” the Mexican War, but commercial expansion was one result. A manifest design had manifest results.

Still, none of this happened overnight. Mexico’s defeat by no means meant that the United States could appropriate a larger share of the Mexican market at will. In the short run, the spike in finished cottons sent to Mexico in 1847/48 did not and could not last (see Figure 3 and Table 5). It reflected the administration of a war tariff by U.S. troops in the occupied ports of Tampico and Veracruz.⁴⁶

Mexico still had commercial weapons, and the demand for protection remained strong. Thus, by the early 1850s, the old complaint was again heard. The United States could expect little from Mexico “whilst the system of prohibitions is observed.”⁴⁷ Meanwhile, the United States brought new territorial pressures to bear during negotiations over the Treaty of the Mesilla in 1853. James Gadsden, the U.S. minister, wanted Sonora and Chihuahua as well but did not get them.⁴⁸ Nor did he get commercial concessions, whatever his original interest in them may have been.

Indeed, after the Treaty of the Mesilla, the Mexican government once again restricted U.S. exports, much as it had done after the Texas rebellion in 1835, and a booming postwar market for U.S. finished cottons all but collapsed (see Table 5). By late 1854, the consul in Veracruz could write, “Santa Anna’s policy destroys commerce, particularly that of the United States.” In a later dispatch he quotes Gadsden, who minced no words. “I had contemplated . . . [directing] the Secretary of State’s attention to the entire Santa Anna Commercial Code—so embarrassing, destructing [sic] and offensive to all trade and intercourse with Mexico—in the hope of convincing him of the necessity of getting rid of the Brigand [Santa Anna].” “Nothing,” concluded U.S. Consul Pickett, “can be more corrupt, false, unequal, and generally pernicious than the entire Mexican commercial system.”⁴⁹

The sources of this vitriol were two: “the” Tariff of 1853 and the Commercial Code of 1854. In practice, four “national” tariffs were in force in 1853, plus state and regional levies in Guadalajara and Monterrey. “How many more there may be in different sections of the country I shall not attempt to record,” wrote the consul in Veracruz. Nor was the Commercial Code much better (or worse). “[O]ne might as well attempt a digest of the laws of the Meades [sic] and Persians or an abridgment of the Chinese Encyclopaedia as a codification of all the imperious arbitrary dicta of the absconded Mexican solon [Santa Anna].”⁵⁰

To the United States, all was chaos. Which tariff applied or to whom duties should be paid was not clear. “Merchants are even now continually imposed on and openly robbed under one or the other of them.” This view was not unique to foreign observers. Mexican historians, too, shudder at the “fiscal disorder” that the declining Santanista dictatorship encouraged.⁵¹

But why should Mexicans educated in the legacies of Guadalupe Hidalgo and the Mesilla assume that open trade with the United States was beneficial? Had it ever been? Its desirability was an axiom only in the minds of U.S. officials, who since the days of Poinsett had repeatedly complained that Mexico impeded trade. James Gadsden was no worse when he concluded, “Let us labor to kill [these barriers to trade] outright and to secure guarantees against their resurrection.”⁵² Mexico’s desire for autonomy (or, indeed, its definition of sovereign interests) figures nowhere in his thoughts.

Santa Anna, on the other hand, had long played cat and mouse with the United States. He understood its interest in enlarging trade and held it at bay as he picked U.S. pockets. “[The] Tariff is not a rigid law in the Republic.. . . His Most Serene Highness [Santa Anna] violates it constantly by selling exclusive privileges,” observed the U.S. consul in Veracruz.⁵³ By misdirecting, stalling, and confounding, Santa Anna’s “chaos” forced merchants to disclose how much they were willing to pay to do business. This may have been dishonest and even inefficient in a broader economic sense, but it was an effective means of extracting rents from U.S. merchants and of restraining their enthusiasm for Mexico. True enough, Santa Anna plundered the state. But his odd ethic was shared by principled idealists whose instincts for personal, political, and national survival were equally indistinguishable. Moreover, Santa Anna displayed a studied ambivalence toward foreign trade. Waddy Thompson, the U.S. minister to Mexico from early 1842 through early 1844, portrayed him as leaning toward autarchy: “[Mexico] had no need of foreign commerce. . . . [It produced] all the necessaries of life.”⁵⁴

The Ayutla movement of 1855 represented, in this context, a shift of substantial proportions. The Tariff of 1856, which permitted the volume of exports to Mexico to grow, was its proximate result. Finished cottons, always a sensitive indicator of the strength of protection, gained ground, and in time their share in exports recovered from the sickening collapse of 1853/54 (see Table 5). Broadcloth, timbers, ready-made clothing, and raw cotton all disappeared from the index of prohibitions. Nominal duties on finished cottons fell by an astounding 70 percent and were lower in 1856 than at any time since 1845. By one estimate, the implicit index of protection on goods from the United States was about 30 percent, an extraordinarily low figure by historical standards.⁵⁵

As a harbinger of Liberal (and liberal) capitalism, the Ayutla movement embraced the national ambitions and possibilities of a bourgeoisie long frustrated by civil unrest. Their notions of growth stressed the expansion of demand rather than the control of supply, a modernizing attitude altogether different from the vague neomercantilism of the later Bourbons and their successors. Witness the words of Guillermo Prieto, who assumed the treasury portfolio in 1855: “The faith I have in free trade is the faith I have in all sublime manifestations of liberty.”⁵⁶

Different, too, was their notion of the political economy of trade. The upswing in exports from the United States that would characterize the third cycle (1867/68-1883/84) marks the end of repeated cycles of annexation and commercial resistance. Better to yield markets than territory, dollars than dominion. Matías Romero put it succinctly: “The best means of impeding annexation is to open the country to the United States . . . with the objective of making annexation unnecessary and even undesirable.” Figures 1 through 3 offer proof that Romero’s observation was historically accurate, even if the United States had no further interest in annexation after the Treaty of the Mesilla.⁵⁷

The “Decade of Civil Wars” (1857-67) was an anomalous one. Contraband distorted normal patterns of trade between Mexico and the United States, and the analysis of corrected statistics is merely somewhat less misleading. Clearly, these years form a bridge (or gap) between the second and third cycles. I shall resume sustained analysis with the Restored Republic (1867-76).

There is, however, a final point. Repeated civil disturbances were disruptive and costly. Foreign conflict may bring prosperity, but enduring domestic crises do not. During the final days of Santa Anna, the U.S. consul in Veracruz noted that “the Pronunciados (Revolutionists) [have] cut off all communication with [Mexico City and] . . . the telegraph has long since been destroyed. . . . The Rebels are determined to seize the public moneys.” So, too, in 1858 with the outbreak of the War of the Reform: “commerce and business [were] completely prostrate, and silver could not be shipped out through Veracruz.”⁵⁸

The costs of remaining on a nearly permanent war footing were severe. Peasants pressed into armed service could not plant or harvest, a major source of disruption to an agrarian economy. Moving armies around the countryside required huge numbers of horses, mules, and oxen to drag artillery and to carry supplies.⁵⁹ Obtaining them from farms, silver mines, and transportation was very costly, and the sample of U.S. import transactions through Laredo, Texas, in Table 6 documents this only too well. Military demands for draft animals during the War of the Reform (1858-60) exhausted the supply of live animals and drove their ordinarily large share of the border trade to zero. This was one of the ways in which persistent instability reduced productivity in nineteenth-century Mexico, and its effects are especially clear from examining patterns of trade with the United States.

The Third Cycle (1867/68-1883/84)

During the third cycle, Mexico’s international position changed significantly. By 1883/84, Mexico had become an important Latin American market for U.S. exports. Real imports (merchandise plus silver) from Mexico grew as well. In 1867/68, they stood at 6.1 million dollars (prices of 1840/41-1844/45). By 1876/77, imports had more than doubled. Indeed, one account suggests that the United States had surpassed Great Britain as Mexico’s principal trade partner by the late 1870s.⁶⁰

The composition of trade changed as well. Mexico began to import capital goods from the United States, and the share of steam engines, sewing machines, machinery, and builders’ hardware grew. Moreover, the share of specie and bullion in imports from Mexico fell to just over 50 percent, while the volume of jute, sisal, and hemp quadrupled between 1871 and 1880. In other words, investment grew, while exports swelled and diversified.⁶¹ Arnaldo Córdova argues that modern capitalist development first appeared during the Restored Republic. Evidence from the trade cycle is consistent with Códova’s argument. It also suggests that Mexico capitalized on the favorable international economic conditions after 1856 as soon as it had attained a reasonable level of governmental stability.⁶²

Yet the beginnings of this expansion were obscure. The Liberals entered Mexico City in the summer of 1867, but for another two years merchants complained that business languished.⁶³ Or was the slump limited to Veracruz? The Free Zone (Zona Libre) along the northern border was a smuggler’s delight, and Matamoros felt no discomfort.⁶⁴ Large shipments of silver to China via San Francisco linked the mines of Zacatecas with the port of Mazatlán.⁶⁵ The old axis of colonial trade and commerce— Mexico City to Veracruz—was neither dying nor dead but saw intimations of mortality. One port’s prostration was another’s prosperity.

Nevertheless, by 1871 some perceived a slow rise in agricultural production, foreign investment, and exports brought by the prospects of peace. In 1872, a new tariff reduced the list of prohibitions and consolidated a number of miscellaneous duties. The proximate result of the revision heartened U.S. cotton manufacturers. Between 1872/73 and 1879/ 80, when the tariff was again revised, the share of finished cottons in U.S. exports rose to more than 13 percent, the highest it had been in twenty-five years (see Table 5), an increase that must have seemed encouraging.

But like all good campaigners, U.S. officials wanted complete victory (or unconditional surrender). For a variety of reasons, they did not get it, at least before 1884. One explanation was the changing terms of trade.

From 1866/70 to 1871/75, the terms of trade with the United States improved by nearly 30 percent as the inflation of the U.S. Civil War subsided (see Table 7). Real silver output grew slowly, but its purchasing power surged. By contrast, from 1871/75 to 1876/80 silver production grew more rapidly, but the terms of trade actually fell. The growth in Mexico’s capacity to import diminished, and the market tightened, as the inflection in the U.S. real export curve in Figure 1 shows. U.S. consular officials blamed shoddy merchandise, inappropriate products, and incorrect packing—all the usual suspects. Nor could the United States finance its exports as the Europeans did, for the United States was still a net importer of capital. But the depreciation of the peso undermined the case for free trade.⁶⁶

Nor did it end there. By 1880, a new tariff and a new regime committed to industrialization were in place. And in view of the rapidly falling freight rates that the railroads brought, the demand for protection necessarily rose, if only to compensate for the fall in the “tariff” of transportation costs. So the Tariff of 1880 became the first in a series of upward revisions that reputedly made Mexico’s nominal rates, particularly on finished cottons, the highest in the world. Since raw cotton was now admitted duty-free, the effective rate of protection, the margin between the cost of imported cotton and the sale price of the finished cloth, must have been substantial.⁶⁷ The industrialists and financiers who had assumed control of the Mexican state knew precisely what they were doing.

These developments—the depreciation of the peso, Mexican industrialization, and renewed demands for protection—signalled a retreat from the Liberal position of the late 1850s and called for new thinking in the United States. It was not long in coming and found expression in the commercial treaty of 1883. As Abram S. Hewitt, a congressman from New York, put it, the treaty, which was ratified but never fully implemented, proceeds from a totally different idea [and] . . . regards Mexico and the United States as integral parts of one commercial system. It is an attempt to establish between the two countries the same condition of affairs that exists between the several States of the Union.”⁶⁸ Indeed, the treaty, which created lists of duty-free goods, was part of a larger congressional movement to establish what a supporter called an “American Zollverein” or customs union embracing all the Americas. But with Mexico the advantages of propinquity were greater, and as Hewitt said, “So long as we get an entering-wedge we ought to be satisfied.” What Poinsett had so long ago essayed, Hewitt and his allies now in part accomplished.⁶⁹

The “entering wedge” came in 1884 with the completion of the rail link between Mexico and the United States. The railroad brought the third cycle to a close and permanently altered economic relations between the two nations. Like Santa Anna, whose death in 1876 coincided with Díaz’s accession to power, Porfirio Díaz would play various foreigners against each other. And like Santa Anna, he would have his victories. But unlike Santa Anna, Díaz no longer had a northern “desert” to mediate between weakness and strength. In fact, during Grover Cleveland’s second term alone (1893-97), the United States sent as much to Mexico as it had during the combined presidencies of Andrew Jackson through Abraham Lincoln. The world had indeed changed.

Reflections on the Balance of Payments

The U.S. balance of payments with Mexico, represented by columns G through I in Table 2, is well worth considering.⁷⁰ The United States consistently ran a visible deficit (column G) that its estimated earnings from shipping services (column H) did not offset. Since the merchandise balance consistently favored the United States, its visible trade deficit was the result of large deficits on the specie balance. By and large, Mexico sent more silver to the United States than required to pay for the goods and services it imported.

Was this unusual, and why did it occur?

A developing country typically runs a deficit on its current account and a surplus on its long-term capital account. Such a country typically consumes more than it produces, especially during economic expansions. To do so, it borrows abroad to finance additional imports. Nevertheless, what is true of the balance of payments in general is not necessarily true of specific cases. Surpluses from one partner may finance deficits with another, for a country’s trade need not be balanced with each of its partners, just balanced overall. Perhaps Mexico achieved its balances partly through trade with the United States. But the evidence is still incomplete.

Reexports yield a more promising clue. If we add them to domestic exports, the cumulative current account balance of the United States with Mexico before the U.S. Civil War (thirty-five years) was approximately 3.8 million dollars or nearly balanced on an annual basis (these calculations are omitted from Table 2). In other words, Mexico did some of its business with Britain, France, and the Hanseatic cities of Hamburg and Bremen through New York. Within reasonable limits for shipping and brokerage, the figures come out about right. Of course, this interpretation is consistent with the commercial rather than industrial character of antebellum U.S. capitalism.

After the U.S. Civil War, Mexican current account surpluses increased, even when reexports are considered. No one factor accounts for them, but several plausible explanations exist.

The first is the export of silver coin to California for reexport to China. Between 1866/67 and 1871/72, the United States reexported 10.6 million dollars of silver specie to China. Over the same period, unexplained exports of Mexican silver to the United States (that is, silver beyond that needed to pay for U.S. exports, reexports, and shipping services) totalled about 14 million. In a crude sense, U.S. reexports of silver to China “explain” about 75 percent of Mexico’s current account surplus, although they abruptly ceased in 1871/72 (there were sizable reexports to Hong Kong in 1873/74).⁷¹

The timing of the surge is no coincidence. In the spring of 1856, the Chinese authorities in the treaty port of Shanghai made the Mexican “dollar legal tender, and the ports of Canton and Foochow soon followed. By early 1858, the demand for Mexican coin in Shanghai was substantial. Once the civil wars in Mexico and the United States were over, there was no obstacle to the remission of Mexican silver to China through California.⁷²

Contraband is another part of the puzzle. By 1878, one U.S. observer wrote that “smuggling [into Mexico] had so largely increased that honest commerce was ruined.”⁷³ Smuggled goods exported by water posed no problem. The items were recorded as conventional exports from the United States’s standpoint. Overland exports to Mexico or Canada, legal or no, were another matter. They went unrecorded until 1893. If significant overland smuggling existed, the exports would go unrecorded on the U.S. side of the ledger.

Yet the specie that paid for smuggled goods necessarily figured as a U.S. import. If overland smuggling were substantial, Mexico would send far more silver than required to pay for recorded exports from the United States. And that is precisely what happened. Overland smuggling from Texas into the northern Free Zone was rampant in the middle and later 1870s. Indeed, it was in the 1870s that smuggling first grew large enough to create noticeable discrepancies in the balance of payments. It thus became a major irritant to bilateral relations, and the existence of the Free Zone itself hung in the balance until 1878.⁷⁴ Mexico’s surpluses of the time—and they were substantial—thus record the unrecorded, if not the unrecordable.

Finally, Porfirio Díaz’s seizure of power was a very near thing, and hardly uncontested. The country was in turmoil, and a sudden fall in the price of silver in late 1878 compounded the atmosphere of “great national calamity.” With reserves apparently short, Treasury Minister Matías Romero was forced to cast about for new sources of revenue, and the U. S. legation reported that he “alarmed all the property and business interests of the country lest [these measures] fall too heavily upon their branches.” Perhaps the nation’s alarmed financial interests took steps to secure their assets abroad. If so, part of the unexplained surge of Mexican silver exports in the late 1870s was capital flight—some silver recorded on current account belongs to the capital account instead.⁷⁵

Conclusions and Implications

Paul Gootenberg and Frank Salford have observed that neither Peru nor Colombia adopted an unrestricted free trade regime much before the 1850s.⁷⁶ To that list we may now add Mexico, at least by evidence of its trade with the United States. Indeed, it was not until midcentury that Mexico adopted an even vaguely liberal commercial policy. But viewed from the perspective of Porfirian Mexico, that opening too was an interlude. With periodic exceptions, Mexico has historically been a high-tariff country.⁷⁷

I have elsewhere suggested a number of factors that made protection an attractive stance in the years before 1840. My purpose here has been to suggest political sources of greater subtlety, generality, and chronological scope. From the Mexican perspective, relations with the United States in the years before 1853 frequently turned on the balance between annexationist pressures and commercial exclusion. In the early national period, the Mexican response to annexationist pressures and losses entailed a considered and defensive effort to restrict trade with the United States. The rise of liberalism in its pure (Juárez) or even manqué (Díaz) forms thus coincided with a commercial settlement of sorts with the United States. And while Santa Anna is no star in the official firmament of Mexican Liberalism, the inchoate form of later tactics to which Sebastián Lerdo de Tejada or Díaz would subscribe is plainly evident in his actions. Díaz, we are told, largely reaped what Juárez sowed. But both learned something from “Su Alteza Serenísima” as well. Even in the “chaotic” years of the early and middle nineteenth century, there were continuities of political style and culture that provide an intelligible intellectual framework.

But Mexican commercial policy ultimately affected economic variables, and it must be judged mainly in those terms. In this context, its efficacy was unquestionable. Despite all that has been said about Mexican smuggling, evasion, and corruption—a litany chanted in the United States with only brief respite since the 1840s—none of these repealed the law of demand.⁷⁸ When Mexican tariffs drove up the relative price of U. S. finished cottons, their consumption within Mexico fell. When prohibitions were laid on U.S. goods, their quantities exported dropped. The evidence is too persuasive to argue otherwise. Even under the most disorganized circumstances, policies that relied on supply and demand for their implementation worked. The spectrum of interpretive possibilities thus offers two alternatives. Either the law of demand is surprisingly robust, or the disorder that reigned in Mexico from the 1830s through the 1870s was institutional rather than administrative. As in most matters, the truth lies between these extremes.

Finally, we must realize that early Mexican regimes were not free to choose. Policies that were politically expedient could also be economically disastrous. No sensible economist could find much to recommend in the whirligig of early national tariff policy. But even a mildly nationalistic (or realistic) one might understand why there was little choice. Tariffs and prohibitions were consistently used to enhance the economic and political stability of the Mexican state, even though these same measures implied a corresponding reduction in national income. Indeed, before 1867, Mexico’s choice may have been between existing poorly and not existing at all. If a condition for long-term growth is the evolution of unfettered, politically autonomous trade, Mexico was, by force of circumstances, unable to meet it.⁷⁹ Mexico’s loss of territory was a burden of U.S. imperialism, but its defensive commercial policy necessarily reduced efficiency and retarded growth as well. Over the course of a century, even small annual losses compound to large disparities in international income levels. Sometimes, little things mean a lot.

Woodrow Borah, John Coatsworth, Albert Fishlow, Pedro Fraile, Stephen Haber, John Huston, Linda Salvucci, Donald Stevens, Barbara Tenenbaum, and two anonymous referees offered suggestions and criticism of this paper. Seminar participants at the University of California, Berkeley, and at the University of Texas, Austin, were helpful as well. An earlier version was presented at the annual meeting of the American Historical Association. The National Endowment for the Humanities and the Social Science Research Council provided financial support.

Correcting Series U321 and U339 for Errors and Omissions

Until 1893, goods carried overland for export were not included in U.S. export totals. This omission affected all overland trade to Mexico. For example, the value of the Santa Fe trade between Missouri and the Provincias Internas is not included in Series U321. According to data supplied by Josiah Gregg in Commerce of the Prairies (1849), the Santa Fe trade averaged nearly $134,000 per year from 1822 through 1844, but yearly swings of up to $100,000 were not unknown.¹ In general, U.S. merchants exchanged dry goods for silver specie and bullion—as much as $180,000 in 1824. This was no small sum. It amounted to more than 7 percent of Mexico’s net exports of silver to the United States in 1824/25.²

Nevertheless, the significance of the trade to the United States is not completely clear. In the 1820s, petitioners from Missouri claimed that the profits of the trade were “an amount considerable in the commerce of an infant state.” Moreover, “the principal article carried to the Internal Provinces is cotton goods, the growth and manufacture of the United States.”³ Yet U.S. cottons were no more than about a third of domestic exports to Mexico in 1825/26, and domestic exports were then only 16 percent of all exports to Mexico. U.S. cottons could not have been more than 5 percent of all U.S. exports (that is, domestic exports plus reexports) to Mexico in the mid-1820s. Unless the Santa Fe trade was unusual, U.S. cottons had little place in it.

Moreover, the marked controversy over granting a “drawback” or rebate on tariffs levied on foreign goods for reexport to Santa Fe suggests that cottons and calicoes from England and France, linens from Germany, and handkerchiefs and stockings from India were staples of the trade. Consequently, we assume that Josiah Gregg’s estimates of the size of the trade should be added to reexports rather than to domestic exports.

The years from 1862 through 1865 also present problems, for both Mexico and the United States were embroiled in civil wars. Conventional political boundaries were blurred, contraband flourished, and there was an unprecedented increase in the value of trade. Why?

From 1851 through 1860, United States imports from Mexico averaged roughly 1 million dollars a year in current prices. By 1865, the figure had swollen to $6 million. Raw cotton alone accounted for $5 million. But Mexico had not suddenly become a major cotton producer. It had become an entrepot (at Matamoros, Tamps., across the Rio Grande from Brownsville, Texas) for raw cotton that could not be shipped from Confederate ports because of the Union blockade. The cotton was then lightered downriver and transferred to ocean-going vessels for shipment to the United Kingdom and to the Union states. To correct the U.S. import figures, we simply deduct the value of cotton imported from Mexico from the import totals and supply the correction in brackets to the right of the “official” figure in Table 2.⁴

U.S. exports to Mexico, Series U339, also increased during the Civil War. Union merchants used Matamoros as an entrepot through which to smuggle supplies to the Confederate states, often under the guise of supplying Juárez and the Liberals in their struggle with the French. As a result, United States (i.e., Union) exports to Mexico between 1862 and 1865 are too large. Annual domestic exports to Mexico in the 1850s averaged $2.3 million. By 1865 they had reached nearly $14 million. For the years 1863 to 1865, I adjusted exports of wheat and wheat flour, cotton manufactures, manufactures of iron and steel, and exports of boots and shoes to Mexico to account for this smuggling. The corrections appear in Table 2 in brackets to the right of the official figures and are based on the same sources used to correct U.S. imports from Mexico.

To verify my adjustments, I estimated a time trend for imports and exports between 1850 and 1870 but omitted 1862 through 1865 as anomalous. I then used the fitted line to predict what imports and exports should have been had the U.S. Civil War not intervened. The “counterfactual” values were not substantially different from the intuitive corrections I provide.

Constructing a Domestic Export Price Index

Readers interested in the nuts and bolts of constructing a base-weighted (Laspeyres) price index should consult Paul Gootenberg, “Carneros y Chuño: Price Levels in Nineteenth-Century Peru,” HAHR, 70:1 (Feb. 1990), 1-56, for a model.

This index is simpler than Gootenberg’s. Four or five goods always dominated domestic exports to Mexico, even though their relative positions changed from year to year. In 1842/43, for instance, finished cottons, raw cotton, and flour accounted for 35 percent of the value of domestic exports. Finished cottons and raw cotton accounted for most exports, and their shares were roughly equal. In 1845/46 they accounted for 55 percent of exports, but raw cotton outweighed finished cottons. In 1852/53 they comprised nearly 80 percent of exports, but finished cottons were most important. How do we handle this?

All price indexes are idealizations, and this one is too. To construct it, I researched the annual “Statements of [Foreign] Commerce and Navigation of the United States” for 1825/26—1858/59; from 1859/60 through 1883/84, I used Commerce of the United States and Other Foreign Countries with Mexico, Central America, the West Indies, and South America. For each year, I determined the five leading exports by value. The weight assigned each was its annual export value divided by the annual value of all domestic exports. After figuring these values for each year, I examined the results to see if any patterns or groupings emerged.

A number of consistencies were evident. For example, in 1825/26-1840/41, finished cottons and wheat flour generally appeared in the top five export goods, although their annual weights varied. To avoid the repetitive calculation of an index whose changing weights required annual “links,” I averaged the annual weights and used the resulting averages as the weights for finished cottons and wheat flour during the subperiod.

For prices, I used the Historical Statistics of the United States (Washington, 1975), Series E 123-134, “Wholesale Prices of Selected Commodities: 1800-1970.” For each year, I multiplied a good’s average weight by its price. To keep other prices constant, I also multiplied everything else (1 minus the sum of the weights of the goods included) by 1. This gives a conservative measure of price change and avoids an implicit exaggeration of the weights that occurs when no other goods are included. The resulting sums (the sum of the weight of each good times its price, plus the “everything else” term) were then divided by the average sum for 1840/41-1844/45. This base period roughly corresponds to the 1839-45 (peak-to-peak) business cycle in the United States and was as normal a time in Mexico’s foreign economic relations as any other. No other period seemed more suitable, or less unsuitable.

Because the composition of trade changed over sixty years, I sometimes adjusted weights, dropped goods that were no longer important, or added new goods. To maintain historical and interpretive continuity, I linked or chained the subperiods by overlapping them and recalculating forward.

The periods and weights used are shown in Table B-1. The weights are rounded to two decimal places and differ a little from the weights I actually used because of rounding.

Some adjustments reflected changes in Mexican policy, e.g., large licensed imports of raw cotton beginning around 1840. Others were needed because of the inflation caused by the Crimean War (1854-56) or the U.S. Civil War (1860-65). Railroad building in Mexico provided the rationale for the construction of the last subperiod.

An important point is that the index of U. S. export prices to Mexico is not the same thing as the index of Mexican import prices from the United States. Distribution and supply in Mexico were always subject to severe interruption, and the operation of international markets was bound to be affected. As a result, the Mexican terms of trade with the United States are not really the reciprocal of the U.S. terms of trade with Mexico. We need four prices, rather than two.

But we do not have Mexican prices, at least not yet. We are forced to use the U.S. export price index as the Mexican import price index. For now, this is an unavoidable simplification.

Estimating the U.S. Current Account Balance with Mexico

The current account balance is the difference between exports and imports of goods and services. The visible trade balance is the difference between exports and imports of merchandise, or net income from merchandise trade. The current account and trade balances differ by the extent of trade in services or “invisibles.” These items include business services (brokerage, shipping, insurance), tourism and emigrant funds, and interest and dividends on foreign investments. The current account measures net income from trade, services, and investment.

Looking at net balances from the U.S. side of the ledger, for our purposes only business services in trade matter. Tourism in either direction did not matter. Mexican migration to the United States before 1897 was small, and even fewer U.S. citizens migrated to Mexico. Interest and dividends on U.S. investment in Mexico could not have been large before the 1890s, or, at the earliest, before the mid-1880s. Wealthy Mexicans had purchased U.S. bonds and equities since the 1830s, but the size of their holdings is unknown. Little specie that Mexico sent north represented a credit to the U.S. current account before the mid-1870s.

But brokerage and the carrying trade were another matter. The United States garnered important earnings from the carrying trade in the early nineteenth century.¹ Data from the annual “Statements of [Foreign] Commerce and Navigation of the United States” in Table C-1 suggest that U.S. vessels controlled the great bulk of the carrying trade between Mexico and the United States. Yet how well can we measure these earnings?

We have several estimates of the cost of importing Mexican silver. In 1842, F. M. Dimond, the U.S. consul at Veracruz, wrote that “all remittances are made in hard dollars the export duty on which is 3.5 percent and freight generally 1 percent. Commission for purchases 2.5 percent and on sale 5 percent and sometimes 8 percent.” He added that insurance on shipments was made in the United States or Europe, there being no institution of that kind in the Republic.”²

The consul in the mid-1850s, J. T. Pickett, also put total costs in the neighborhood of 10 percent.³ But the meaning and apportionment of these costs depends on the definition (and nationality) of the buyer, the seller, and the broker. To avoid interminable complications (and calculations), I assume that citizens of the United States buying silver produced a net charge against the United States of 1.5 percent (1 percent freight payable to U.S. shipping less 2.5 percent commission on purchase payable to a Mexican broker). There are obviously other ways of looking at the matter, but this procedure is a conservative reading of limited evidence. The nature of the silver market, of course, goes well beyond the scope of this essay.

Mexico also exported logwood, cochineal, dyestuffs, hides, and other commodities to the United States. But we have no basis for estimating U.S. earnings on them and must omit them from our calculations. Specie mattered most anyway.

U.S. earnings on exports to Mexico were composed of earnings on domestic exports and earnings on reexports. A report of 1885 concluded that commissions, insurance, and freight from the United States to Mexico added 20 percent of the value of goods.⁴ In short, we multiply the value of exports plus reexports by 20 percent and credit the result to the United States. We could refine the estimate by adjusting the yearly totals by the proportion of cargo carried by U.S. vessels, but that gives too great an impression of precision. These are orders of magnitude and should be understood as such.⁵

Texas and Changing Patterns of Trade

I suggested that the Texas rebellion produced a dramatic change in the pattern of trade between Mexico, the United States, and Great Britain. Here I wish to demonstrate precisely how pronounced that shift was.

In order to measure the shift in trade patterns “caused” by affairs in Texas and by the Mexican reaction to them, I converted U.S. domestic exports and all British exports to Mexico to index numbers whose base year is 1835 (i.e., 1835 = 100). I then examined the behavior of both series to see if the pattern of export behavior changed in any significant way after 1835. In analyzing time series, it is customary to remove the underlying trend by differencing (i.e., subtracting successive observations from each other). In this case, I examined both first differences (the rate of change of the index numbers) and second differences (how much the change itself varied). While a distinct pattern was evident in both the first and second differences of the index numbers, the series of second differences shows the pattern most clearly.

Figure D-1 demonstrates this graphically. Before 1835, U.S. and British exports to Mexico move in more or less parallel fashion, although the swings in U.S. exports (the short dashed line) are less pronounced than those in British exports (the solid line). But around 1835/36 (where the vertical axis serves as a kind of before-and-after signpost), a new pattern emerges. Growth in British exports is now accompanied by a slowdown in U.S. exports, and vice versa. Graphically, the pattern of trade no longer appears as parallel fluctuations driven uniformly by demand, but as reversed peaks and troughs. The altered trade flows continue until the outbreak of the Mexican War.

There can be no clearer demonstration that around 1835 there was a dramatic change in trade relations between Great Britain, the United States, and Mexico. Mexico’s markets were thereafter “opened” to one party mostly at the expense of the other, something not much in evidence before. In other words, Mexico now pitted one partner against the other in the contest for markets.