2. The almojarifazgo de indias and its juro obligations

The almojarifazgo rates remained constant on both sides of the Atlantic during the 17^th century. From 1566 right up to the early 18^th century, it was levied at 5 per cent on leaving Spain and 10 per cent on entering the Indies. On the return leg, it was 2.5 per cent on departure (though it appears not to have been levied when leaving Portobelo) and 5 per cent (plus 10 per cent of alcabala, the main Castilian sales tax) on entering the western Andalusian ports (Antúnez Acevedo, Reference Antúnez Acevedo1797, pp. 208-214; Chaunu and Chaunu, Reference Chaunu and Chaunu1955-1960, I, p. 251).

Although the Spanish and American almojarifazgos shared a name, they differed in more than their rates. On the one hand, whereas in the Indies the tax was always administered by royal officials, in Seville and Cadiz it was managed by tax farmers in 85 of the years between 1595 and 1700 (Pulido Bueno, Reference Pulido Bueno1993, p. 118). On the other hand, in Spain its revenue (administration costs deducted) was formally earmarked to service juros, which typically yielded between 5 and 7.14 per cent of the capital lent to the king. When the tax was contracted out, juristas received their yield from the tax-farming rent. I discuss the relationship between the Crown's policy and the revenue of the almojarifazgo de Indias in section 4. Here I provide the history and main characteristics of the duty.

The 17^th century saw the gradual demise of all Castilian juros. All domestic revenues (whether royal or municipal) were severely affected by the economic and demographic crisis that hit Castile in the first half of the century. The service of the long-term debt suffered correspondingly, although the fate of the juros depended on the specific tax to which they were pledged. To make things worse, fiscal pressure kept mounting in Castile. Desperate to find the resources to finance military campaigns in Flanders and elsewhere, the Spanish Crown renewed and created taxes (millones and cientos), increased rates and kept resorting to short-term loans and issuing juros, among other expedients. By 1630, the servicing of juros (called the situado) absorbed most of the Crown's ordinary revenue.Footnote ⁴ In 1635, as it had done before in 1625 and 1629, the Crown began confiscating a percentage of the yield of juros—usually between 35 and 65 per cent. These confiscations, known as valimientos and descuentos, continued for the rest of the century (Toboso Sánchez, Reference Toboso Sánchez1987, pp. 174-184). As a result, the reputation of the monarchy's long-term debt was seriously compromised. The juristas of the almojarifazgo de Indias endured all these problems and then some, for no other ordinary revenue collapsed in the way the main transatlantic trade tax did.

In 1599, of all the Royal Treasury's panoply of domestic revenues, the almojarifazgo de Indias and the almojarifazgo Mayor de Sevilla (Andalusia's main duty on non-colonial external trade) were among the most substantial; the former accounted for 2.22 per cent of the total revenue, the latter for 5 per cent. When Spain's customs revenue is considered, they made up 15.2 and 34.3 per cent, respectively (Andrés Ucendo and Lanza García, Reference Andrés Ucendo and Lanza García2008, pp. 153-155). Even though the two taxes were contracted out, by the 1600s the Royal Treasury barely received anything, as the tax-farming rents went almost entirely to service juros. A 1607 report, prone to hyperbolic language, stated that the yields from the juros on the two Andalusian almojarifazgos sustained «the whole of Castile, [from] the most important men [to] the neediest people and all the monasteries» (Gelabert, Reference Gelabert1997, pp. 127-128). However, the vagaries of colonial trade meant that such yields could not be guaranteed every year. As early as February 1590, the Casa de la Contratación (the House of Trade, the regulatory institution for transatlantic private enterprise, pursuing the state's fiscal objectives) reported the impossibility of meeting all juro obligations. The following year, as a moratorium was requested, the Consejo de Hacienda (Council of Finance) wrote secretly to the Seville authorities (the juros were serviced at the city level to ensure their credibility) ordering that the local treasurers, «with dissimulation, delay the payments of the juros» (Ulloa, Reference Ulloa1977, p. 294). Tempers cooled with the good fiscal results of the following years. However, in the 1600s, the so-called no cabimiento (non-payment of a varying number of juros due to insufficient tax revenues) of the almojarifazgo de Indias's juro obligations would cease to be an exception, instead becoming the norm for the rest of the century. And, crucially, the Crown was not accountable for it. Since the possibility of non-payment was already part of the original loan contract, «when specific tax revenues dried up, it did not constitute bankruptcy to fail to service juros» (Drelichman and Voth, Reference Drelichman and Voth2014, p. 91).

As plotted in Figure 1, the revenue of the almojarifazgo de Indias went through three distinct phases between 1580 and 1700. Up to 1633, although with significant swings, it averaged 83.5 million maravedís a year; from 1634 to 1660, 28.6 million a year; and from 1661 to 1700, merely 6.5 million a year. From the first to the third period, therefore, the annual average revenue shrank by 92 per cent. Overall, the tax-farming yearly rent reflected the drop in revenue. Still, revenue and tax-farming rent did not always evolve in parallel. When the revenue began to fall in earnest in the second half of the 1630s, the tax farmers found themselves between a rock and hard place. At the time, they were a company of Portuguese businessmen led by the conversos Marcos Fernández Monsanto and Felipe Fernández Dorta, who had taken on the management of the two Andalusian almojarifazgos in 1632. The company was contractually obligated to pay the Crown 293 million maravedís a year for the two taxes (87.5 million for the almojarifazgo de Indias), although this amount was lowered by 20 million when war broke out against France in 1635. Their period in charge lasted 13 years and was marked by all sorts of problems. Fewer and fewer juristas were paid and the Seville authorities were inundated with complaints. The contract (arrendamiento) was unilaterally terminated by the Crown in 1644, as the Portuguese company had by then gone bankrupt, owing juristas 423 million maravedís (Domínguez Ortiz, Reference Domínguez Ortiz1983, pp. 203-205; Pulido Bueno, Reference Pulido Bueno1993, pp. 123-129). The tax revenue shrank even further in the following years, partly due to the war with England. After 1660, the tax-farming rent followed the revenue closely at century-minimum values. By then, the almojarifazgo de Indias and almojarifazgo Mayor de Sevilla were always contracted out to the same tax farmer.

Figure 1. Revenue of the almojarifazgo de indias and its juro obligations, 1580-1700 (current values, million maravedís).

Sources: Archivo General de Simancas (hereafter AGS), Contadurías Generales (hereafter CG), 1169, 1175, 1181, 1203, 1208, 1213, 1218, 2921, 2923, 2948 and 2955; «Contrato del capitán Juan González de Guzmán», Seville, 3 January 1603, Biblioteca Nacional de España (hereafter BNE), Mss/8553, ff. 82r-101v; «Carta de recudimiento», Madrid, 6 March 1699, BNE, VE/192/95; Alloza Aparicio (Reference Alloza Aparicio2019, p. 204); Lorenzo Sanz (Reference Lorenzo Sanz1979-1980, pp. 409-410 and 414-415); Pulido Bueno (Reference Pulido Bueno1993, pp. 126 and 143-144); Toboso Sánchez (Reference Toboso Sánchez1987, p. 280).

Figure 1 plots a third crucial element: the yearly service of all the juros issued against the almojarifazgo de Indias. The information available is patchy but the trend is unequivocal. In the latter years of the 16^th century, from 1580 to 1598, the yearly service ranged between 52 and 78 million maravedís a year, usually well below the tax revenue. The yearly service did not grow during the reign of Philip III (r. 1598-1621) because no juros were issued then—nor were the existing ones redeemed. Things changed dramatically with Philip IV (r. 1621-1665). Although there are no data available to plot the precise evolution of the yearly service, by the beginning of Charles II's reign (r. 1665-1700) it had risen to 137.5 million maravedís (or over half a million pesos), where it remained until 1714 (Toboso Sánchez, Reference Toboso Sánchez1987, p. 173).Footnote ⁵ In other words, by the last third of the 17^th century (and probably by the 1640s), the no cabimiento reached gigantic proportions; between 95 and 96.6 per cent of the annual service was left unpaid.Footnote ⁶

Such a collapse raises the question of which juros were paid and on what basis. Anywhere in Spain, a local order of seniority, called antelación, ranked the juros according to their date of issuance. When the revenue was enough to cover all interest payments, the order of seniority was of little relevance. However, whenever the revenue became insufficient, antelación determined who would be paid and who would not, as only those juros covered by the net revenue or the tax-farming rent would be serviced (Álvarez Nogal, Reference Álvarez Nogal2009, pp. 47-49). There was also another criterion for establishing an order of priority. The Crown always tried to safeguard the juros owed to ecclesiastical institutions such as convents, monasteries, chaplaincies, brotherhoods and hospitals by granting them the status of reservado, which exempted the annuities from full and partial confiscations. By 1665, there were 1,307 juros on the almojarifazgo de Indias in circulation; 103 of them (7.9 per cent), with a combined yield of 7.4 million maravedís (5.4 per cent), belonged to ecclesiastical institutions (Toboso Sánchez, Reference Toboso Sánchez1987, pp. 207-208). However, not all of them were paid in the final decades of the century. In 1700, for instance, the treasurer of the almojarifazgo made 150 payments to 42 juros (ownership of a juro could be divided between several juristas) for a total amount of 3,131,517 maravedís—merely 2.3 per cent of the yearly service.Footnote ⁷

Juristas could do little once the revenue stream began to fall. When faced with no cabimiento, the most powerful and better-connected of them often petitioned the Crown to move (mudar) their juros to a more reliable revenue stream. However, those whose wishes were granted were few and far between (Marcos Martín, Reference Marcos Martín2017, pp. 556-562). This is particularly true in the case of the almojarifazgo de Indias. Permanently pledging their juros to another tax appears to have been an overambitious request. Instead, a small number of petitioners managed to get some arrears paid by the treasurers of other taxes.Footnote ⁸

3. The evolution of trade with Spanish America

One possible reason for the collapse in the revenue of the almojarifazgo de Indias is that the transatlantic trade itself contracted similarly. In this case, the fall in tax revenue would have been the direct consequence of a long-running trade crisis. Did such a crisis occur? Although the quantification of the commercial exchanges between Spain and Spanish America poses problems that cannot be resolved, it appears that transformation, rather than crisis, is the term that best captures what happened in the 17^th century.

The Chaunus (Reference Chaunu and Chaunu1955-1960) were the first to delineate the major cycles of Spain's transatlantic trade up to 1650. They looked at traffic volume, the number of ships and the amount of precious metals transported to Spain. According to their work, after an expansive cycle that began in the 1560s, the Carrera reached its zenith from 1580 to 1592, stagnated from 1593 to 1622 and began to contract from 1623 onwards. How long did this contraction last? It was long thought that the decline worsened after 1650 (Hamilton, Reference Hamilton1934; Chaunu and Chaunu, Reference Chaunu and Chaunu1955-1960; Domínguez Ortiz, Reference Domínguez Ortiz1969). This view seemed to make perfect sense. After all, the commercial fleets, which in theory should have departed every year, not only became less frequent (setting off every two or three years) but also decreased in size over time. What, though, was the extent of the contraction? García Fuentes (Reference García Fuentes1980) was the first to provide precise data series for the second half of the century. As explained below, his figures for registered bullion (and the rest of colonial goods) are of little use since registration of goods for the fleets returning from the Indies ceased to be compulsory from 1660 onwards.Footnote ⁹ On the other hand, his data for merchant vessels and shipping tonnage are more useful.Footnote ¹⁰ Oddly enough, García Fuentes (Reference García Fuentes1980, p. 223) chose not to link his tonnage series with those of the Chaunus, arguing that the French historians' data comprise warships, slavers and merchant ships without distinction. However, this is not entirely correct. It is indeed the case that the Chaunus do not distinguish between types of ships on the return leg, but on the outward leg the merchant ships are duly differentiated. García-Baquero (Reference García-Baquero, Thompson and Yun Casalilla1994) realised García Fuentes's lapse and managed to link the ships and tonnage series of the two halves of the century. Still, he did so by including both the outbound and returning vessels, effectively overlooking the fact that the Chaunus include men-of-war along with merchant ships in their return-leg lists while García Fuentes does not. As presented by García-Baquero (Reference García-Baquero, Thompson and Yun Casalilla1994, p. 117), the century-long downward trend seems nothing short of catastrophic; by the 1700s, 90.7 per cent of the 1610s tonnage had been lost. However, this calls for some important refining.

First, the slump appears less severe if we use the 1690s instead of the 1700s (which were marked by war, disrupting transatlantic communications) as the final decade. Moreover, to offer a more reliable picture of the evolution of the trade volume, it is better to use the outbound tonnage figures only. There are two reasons for this: on the one hand, as mentioned above, the Chaunus only distinguish between types of ships on the outbound leg; on the other, García Fuentes's data on ships returning to Spain are incomplete. When registration of goods ceased to be compulsory on the fleets' return leg, the official number of vessels sailing back to Cadiz and Seville dropped significantly. This is a fact to which neither García Fuentes nor García-Baquero pays any attention. Whereas the return to outward tonnage ratio had been 0.78 in 1621-1650 (Chaunu and Chaunu, Reference Chaunu and Chaunu1955-1960, V-1, pp. 349 and 352), that of the last three decades of the 17^th century was only 0.34. It is not by chance that the ratio began its downward trend in the 1660s. In the 1690s, it bottomed out at a mere 0.21. Presumably, this was not because four out of five vessels chose not to return to Spain but because nothing they carried on their way back was registered, and therefore, upon arrival, the ship masters had no registro to hand over to the House of Trade (the registros were García Fuentes's primary source).

Unfortunately, concentrating solely on the outbound tonnage is a far from ideal approach because the Chaunus included merchant ships that supposedly sailed to the Indies unofficially, whereas García Fuentes focused strictly on official figures, downplaying the effect of unauthorised ventures. That the figures for the two halves of the century are as much the product of the historians' archival work as of the actual vagaries of trade is strongly suggested by their estimates for 1650, the only year when the two data series overlap. For this year, the Chaunus report the departure of 40 merchant ships with a total capacity of 9,238 tons. García Fuentes reports the same number of vessels but puts their total capacity at 6,108 tons. Unfortunately, we cannot delve into the latter figure because García Fuentes does not provide individual details of ships, but the Chaunus do. I have double-checked their data at the Archivo General de Indias in Seville. Three of their 40 ships (1,400 tons in total) did not depart in 1650 but in 1644, 1647 and 1648, which means the Chaunus recorded them twice,Footnote ¹¹ and there are two vessels that I could not locate. However, I came across the registros of four additional ships (690 tons in total) that the Chaunus missed out.Footnote ¹² To sum up, at the Archivo General de Indias I found 39 ships for 1650, with a total capacity of 8,283 tons—below the Chaunus' figure but well above that of García Fuentes.Footnote ¹³ Nevertheless, no substantial conclusion can be drawn from a single year. Despite their deficiencies, I must make use of the data series provided by the above historians. By my estimate, of the total tonnage carried to the colonies in the 17^th century, 25.7 per cent was sent after 1650; and from the 1610s to the 1690s, the tonnage loss amounted to 76.9 per cent (rather than the 90.7 per cent drop reported by García-Baquero for the 1700s). This is still a very significant slump, but one that may accommodate better the other part of the puzzle: the value of the cargoes.

Since customs officials were not allowed to open the bales, bundles and crates containing the most important cargo shipped to the Indies, such as textile products (more on this in sections 4 and 5), the only way to assess the value of the Spanish exports is by looking at how much silver was shipped in return.Footnote ¹⁴ After all, it is well established that in the 16^th and 17^th centuries, bullion accounted for over 80 per cent of the value of Spain's colonial imports (Chaunu and Chaunu, Reference Chaunu and Chaunu1955-1960, VI-1, p. 474) and that returns from commercial transactions provided a similar percentage of the remittances (Oliva Melgar, Reference Oliva Melgar2004, p. 226). Did the treasure shipped to Spain suffer a contraction similar to that of the tonnage over the course of the 17^th century? This is where the idea of a century-long crisis is turned on its head.

Initially, the historians contending that the colonial trade had shrunk reached their verdict by looking at the official bullion imports to Spain, which went down from 52 million pesos in 1606-1610 to 5.5 million in 1656-1660 (Hamilton, Reference Hamilton1934, p. 34). Delving into the official figures produced after 1660 was rightly considered of little use, but there was a broad consensus that if the registration of colonial goods had remained compulsory after that year, things would have looked equally bleak, if not worse. However, there were alternative sources that radically contradicted that gloomy view.

The first to raise a cautionary flag was Girard (Reference Girard2006) in 1932. He looked at the reports sent to Paris by the French consul in Cadiz, Pierre Catalan. Catalan spoke of imports of bullion worth between 17.4 and 32 million pesos per fleet during the reign of Charles II. Sadly, Girard's work was ignored for a long time. Only in the 1970s and 1980s was there a resurgence of the view that the Carrera did not suffer a century-long crisis (Morineau, Reference Morineau1969; Everaert, Reference Everaert1973; Kamen, Reference Kamen1980; Fernández de Pinedo, Reference Fernández de Pinedo, Pérez Picazo, Lemeunier and Segura1986). Other reports by foreign consuls, along with merchants' correspondence and, notably, Dutch gazettes of the time, confirmed that the commercial fleets brought previously unheard-of riches from the Indies in the second half of the 17^th century. Morineau (Reference Morineau1985, p. 256), the most vocal advocate of the alternative sources, claimed that the alleged crisis of the Carrera was «a kind of collective hallucination». He constructed a series of the precious metals sent to Spain during the entire century (indeed, during the entire colonial period), replacing the image of decadence with one of a spectacular boom. His five-year totals for 1660–1700 dwarfed those of the first decades of the century. More recently, Oliva Melgar (Reference Oliva Melgar, Lobato Franco and Oliva Melgar2013) reached a similar conclusion using the correspondence sent to Florence by the Tuscan diplomatic representatives in Madrid. They had access to the reports that the Spanish Crown received from Seville and Cadiz following the arrival of each fleet. Even though the Tuscans' data are incomplete (five fleets are missing), the message is clear enough. Just like Morineau, Oliva Melgar (Reference Oliva Melgar, Lobato Franco and Oliva Melgar2013, p. 154) dispels any notion of crisis, insisting that, as the 17^th century progressed, «more silver than ever before was arriving [in Spain]».

Figure 2 sums up the data on both the outbound tonnage to the Indies and the bullion remittances to Spain discussed above. The divergent trends pose a seemingly intractable conundrum. How was it possible that the largest quantities of bullion arrived in Spain just when trade volume had shrunk to its century minimum?Footnote ¹⁵ We may begin by wondering how reliable the figures provided by the consular reports and Dutch gazettes are. Macleod (Reference MacLeod1989, pp. 166-167) suspects that Morineau attributes «too much accuracy to the Dutch gazettes». To an extent, this may well be so. Recent silver-production data strongly suggest that the Dutch gazettes exaggerate. TePaske and Brown (Reference TePaske and Brown2010, pp. 311-312) find that the figures reported by the gazettes surpassed the registered silver output at the American mines by 33 per cent between 1660 and 1710. However, they also admit that the rate of fraud at the mines in the same period may well have ranged from 25 to 50 per cent. If other (less important) destinations of the silver are factored in—the Philippines via Acapulco, contraband with foreign interlopers in the Caribbean and Buenos Aires, and the monetary needs of the expanding Spanish American economy—then it becomes apparent that the data published by Morineau, Oliva Melgar and others are likely to be off the mark. After all, most figures are suspiciously rounded-up and, perhaps more worrying, differ between gazettes and consuls and from consul to consul.Footnote ¹⁶

Figure 2. Merchant shipping tonnage to the Indies and imports of American bullion to Spain, 1580-1700.

Sources: Chaunu and Chaunu (Reference Chaunu and Chaunu1955-1960, VI-1, pp. 384-391); García Fuentes (Reference García Fuentes1980, pp. 224-225); Morineau (Reference Morineau1985, p. 250); Oliva Melgar (Reference Oliva Melgar, Lobato Franco and Oliva Melgar2013, p. 151-152).

However, although the upper boundary is likely to have been below Morineau's figures, the lower boundary can be established with more confidence. Another additional source, this one official, confirms that very substantial amounts of bullion arrived in Andalusia: the royal mint of Seville coined just over 350 million silver pesos between 1650 and 1700 (Pérez Sindreu, Reference Pérez Sindreu1992, pp. 256-258), that is, 53.4 per cent of the figure provided by Morineau for the same period. If we consider that large quantities of bullion brought by the fleets were transhipped to foreign vessels upon arrival, without ever touching Spanish soil, then the vast magnitude of the American remittances becomes apparent. The idea of a gradual decline of the Carrera must therefore be discarded.

How can the drop in tonnage be accounted for? The Chaunus (Reference Chaunu and Chaunu1955-1960, I, p. 146) insist that «tonnage is […] the best possible quantitative approximation to the economic life of the Spanish-American Atlantic area». This, however, seems unconvincing in light of what happened with the bullion remittances. Something else was going on. Morineau (Reference Morineau1985, p. 252) suggests there was a major transformation in the nature of the cargoes; as the century advanced, the fleets carried «more velvet from Genoa and bayettes from Colchester, cloth from Brittany and lace from Bruges, less oil and less wine»—that is, fewer cheap, bulky cargoes and more expensive, low-volume goods. This would appear to be consistent with another crucial development: the growing self-sufficiency of the colonial economy in the 17^th century (Assadourian, Reference Assadourian1982; Klein and Serrano Hernández, Reference Klein and Serrano Hernández2019). However, García-Baquero (Reference García-Baquero, Thompson and Yun Casalilla1994, p. 123) is not persuaded by Morineau's argument (which admittedly the French historian does not develop in depth), contending that such a radical transformation in the quality of goods exported to the Indies is uncertain. For him, only contraband (within the official system) could account for the shipment of large amounts of silver.

However, the transformation in the quality of exports did take place. Writing to the king in 1622, Rodrigo de Vivero y Velasco, president of the Audiencia of Panama, explained that of all the fleets that had ever arrived at the isthmus, it was the 1592 fleet that had paid the most taxes, 708,088 pesos. In stark contrast, he noted that the taxation of the fleet arriving in 1620 had yielded only 60,444 pesos. In his opinion, the drop in revenue could hardly have been the result of a trade contraction as he was convinced that the 1620 fleet had brought a cargo as rich as that of 28 years before. Vivero y Velasco explained that the 1592 fleet was made up of

The variation in the quality of the goods exported to the Indies will be further examined in the following sections. Before moving forward, however, we must ask a vital question: if the value of Spain's transatlantic trade did not endure a long-lasting crisis in the 17^th century, how can we explain the fact that the almojarifazgo de Indias, an ad valorem duty, collapsed? For Vivero y Velasco and most historians, fraud was the sole culprit. Was this the case?

4. Taxation and fraud before 1660

Any serious attempt to examine the fiscal organisation of the Carrera de Indias and the fraud associated with it must start by making two crucial distinctions that are often overlooked. First, it is essential to distinguish between the outbound and return legs, since taxation encountered entirely different problems and followed contrasting procedures depending on whether the fleets were sailing to the Indies or returning to Spain. Gauging the value of the so-called ropas—the generic word for clothing and other textile products, which made up the lion's share of the outbound cargoes—was a serious problem for customs officials. In contrast, none of the silver officially arriving in Andalusia went through customs. Instead, before 1660, the precious metals were taken directly to the Treasure Chamber (Sala del Tesoro) in the Alcázar of Seville (the headquarters of the House of Trade) and held there for two or three months until the avería was deducted (Lorenzo Sanz, Reference Lorenzo Sanz1979-1980, II, pp. 131-132). The second distinction is between the two main duties, the almojarifazgo and the avería, because their burden did not fall equally on all the goods traded. The almojarifazgo primarily targeted Spain's exports to the Indies. It was also applied to imports from America, with the crucial exception of precious metals.Footnote ¹⁸ The avería, on the other hand, was applied to Spanish exports, at rate of 1 per cent on departure, from the first decade of the 17^th century onwards (Chaunu and Chaunu Reference Chaunu and Chaunu1955-1960, VI-1, p. 416). The bulk of its burden fell on the return cargo, including—and above all—precious metals. Before 1660, the avería's return rate, which was entirely levied upon arrival in Seville, varied depending on the final bill for the costs of providing military defence for the fleets, and the official amount of treasure carried to Spain. As will be explained below, the drop in the official silver imports, the consequence of massive fraud in the 1640s and 1650s, forced the 1660 reform.

In the 16^th century, determining the tax base of the exports to the Indies (known as avalúo de mercancías) was already giving rise to heated disputes between customs officials and merchants. Rather than clarify things, the Crown added fuel to the fire. A 1524 royal decree exempted merchants from opening bales, crates and other packages at customs, a privilege that would be upheld until 1778. In the second half of the 16^th century, merchants had to present the invoice (encaxe or factura) of the merchandise they intended to ship (Chaunu and Chaunu, Reference Chaunu and Chaunu1955-1960, I, pp. 92-96). Fearing that merchants were under-reporting the value of their goods, tax farmers occasionally demanded sworn declarations along with the invoices, but the House of Trade (representing the Crown) blocked such attempts. In August 1586, Philip II (r. 1556-1598) issued a royal decree barring customs officials and tax farmers from demanding oaths from merchants (Antúnez Acevedo, Reference Antúnez Acevedo1797, Appendix X). However, the quarrels at the customs houses resurfaced from 1596 to 1604. Every attempt the tax farmers made to determine the market price of the merchandise was met by merchants' refusal to load their goods onto the fleet ships. The House of Trade would then order the tax farmers to back down so that «the fleet's dispatch was not impeded».Footnote ¹⁹ The problem persisted in the early decades of the 17^th century, as evidenced by further decrees, issued in 1609, 1617, 1623 and 1629, insisting that asking for sworn declarations or inspecting the content of the packages was illegal (Veitia Linaje, Reference Veitia Linaje1672, I, p. 118; Bernal, Reference Bernal1992, p. 216). To put the disputes to rest, Philip IV (r. 1621-1665) enacted a new method of tax assessment in January 1629. From then on, bales and other packages containing textile products were to be assessed not by the declared value of their contents but by their weight. Each bale (fardo or frangote) would be valued at 5,100 silver maravedís per arroba (11.5 kilograms) if it was to be shipped to the Isthmus of Panama and at 3,600 silver maravedís per arroba if its destination was Veracruz (Veitia Linaje, Reference Veitia Linaje1672, I, p. 119). The appraisal amount would serve as the principal on which to apply the duties. Due to resistance from tax farmers, the new assessment method was not implemented immediately. The ships' goods registries (registros de mercancías), which recorded all the goods shipped and their customs value individually, show that appraisal by weight was only instituted in 1635.Footnote ²⁰ Under-reporting of value ceased to be a form of fraud from that moment onwards. Fiscal pressure was thus officially lowered. Surprisingly enough, the profound implications of the new appraisal method remain unexplored.Footnote ²¹

Only two historians have paid attention to the drop in revenue from the almojarifazgo de Indias after the 1630s, which is plotted in Figure 1. Pulido Bueno (Reference Pulido Bueno1993, p. 146) considers that it was the direct consequence of «the decline and discontinuity of Spain's trade with America». In contrast, Alloza Aparicio (Reference Alloza Aparicio2019, p. 203) takes the view that the tax revenue could hardly reflect the evolution of trade due to the massive increase in fraud. Why, though, would fraud on the outbound leg be higher in the middle decades of the century than previously? The incentives were certainly not there; outbound cargoes were never subjected to embargoes and, as explained above, ordinary fiscal pressure was lowered in 1635. Moreover, by order of the Council of the Indies, the visitador Juan de Góngora initiated an inspection of the House of Trade in 1642 that would last several years. His findings showed that there was indeed fraud, and that the fleet generals and admirals and the House of Trade officials were involved in all sorts of illegal activities and corruption (Heredía López, Reference Heredia López2021). However, the almojarifazgo de Indias does not appear to have been their primary victim, or at least not to an extent that differed from previous decades. So, if fraud was not the leading cause of the collapse of the main customs duty, what was? I contend that the drop in revenue was primarily (though not only) caused by the new appraisal method. This is for two reasons. First, because the customs value per arroba was set particularly low. Secondly, because appraisal by weight accelerated the transformation in the quality of goods exported to the Indies, encouraging merchants to load more valuable, less bulky merchandise.

The approval of a measure that so blatantly harmed customs revenue raises the question of why the Crown would do such a thing. Unfortunately, I have yet to find a document where this policy was explicitly laid out. However, my interpretation is the most plausible explanation for a seemingly puzzling decision. Indeed, once the monarchy's long-term domestic debt is brought into the equation, it makes more sense. Increasing fiscal pressure on the customs houses would not have benefited the Royal Treasury but instead the juristas, and it would have delayed the departure of the fleets. This is not to say that fraud became unimportant now that ordinary fiscal pressure had been officially lowered. On the contrary, fraud became widespread in the mid-17^th century, but it was concentrated mainly on precious metals and affected the avería's return rate. To a large extent, it was a response to the predatory actions carried out by the Crown.

While enacting measures that eroded the almojarifazgo de Indias, the Crown began to apply fiscal pressure by means that circumvented the angry juristas and brought silver straight to the Royal Treasury's coffers. It was a trade-off that went too far. Waging war in Flanders and elsewhere demanded substantial financial resources that were extremely difficult to assemble. The military campaigns were funded by increased fiscal pressure on Castile and recourse to short-term loans, with new emissions of juros and the king's American treasure as collateral. However, when these sources of funding were deemed insufficient (as they often were), the situation called for even more desperate measures. In this context, the return of a fleet presented a tantalising possibility, one that was greatly facilitated by the institutional set-up: seizing the silver held in the Treasure Chamber belonging to private individuals (primarily merchants). Such an extreme measure allowed for imposing loans in favourable conditions (which, to add insult to injury, were seldom reimbursed), bartering silver for juros or vellón (mixed silver and copper coinage that went through multiple debasements), and forcing extraordinary contributions in the form of indultos (pardons). It was especially from 1620 to 1660 that silver embargoes (or the threat of them) were most frequent and damaging. Merchants responded by avoiding silver registration. According to Morineau (Reference Morineau1985, p. 22), fraud on the return leg hovered around 15 per cent in the 1620s and regularly exceeded 50 per cent thereafter. In 1641, 1642 and 1646, it was above 80 per cent. In 1659, the last year of his series, it was 66 per cent. Naturally, the increase in fraud led to an upward spiral in the avería's return rate, provoking the so-called «crisis of the avería». From 1600 to 1627 it had stayed at 6 per cent, but from 1628 onwards it shot up. In 1630 it rose to 31.5 per cent; in 1632 to 35.5 per cent. In June 1644, in an ultimately futile attempt to alleviate the situation, Philip IV decreed that private individuals should pay no more than 12 per cent; the extra cost would be borne by the Royal Treasury. However, the Crown's credibility had crumbled by then, and the measure did not reduce the tax rate or fraud (Álvarez Nogal, Reference Álvarez Nogal and Armillas Vicente1998). In the early 1650s, most silver arriving in Spain was off the books. In 1653 the avería rate reached a record 99 per cent; in 1656 it was 49 per cent (Veitia Linaje, Reference Veitia Linaje1672, I, p. 157).

Reversing this critical situation required bold, persuasive measures. First and foremost, the Crown had to convince the merchants that it would no longer interfere arbitrarily in their business. The very survival of the Carrera de Indias system depended upon it.

5. Taxation and fraud after 1660

The final solution to the avería crisis took several years to crystallise. Although the Crown would have liked to impose its will on the merchants at all times, the truth is that predatory actions such as silver embargoes were the exception rather than the rule. It was far more common for negotiations to take place between the Crown and the consulados on both sides of the Atlantic (those of Seville, Mexico City and Lima), although they seldom went smoothly.

In December 1654, the Lima Consulado explained to the king (as if he needed any explanation) why the transatlantic trade and remittances of registered silver to Spain had shrunk and why merchants preferred to risk losing their silver by shipping it illegally in the form of piñas (pinecone-shaped ingots) and barretones (bars) and evading taxes. For the Peruvians, the silver embargoes initiated in Seville in 1620 and the rise in the avería's return rate were the primary reasons. In their view, two measures could reverse the situation. First, under no circumstances should the Crown take the registered silver belonging to merchants and other private individuals. Secondly, 12 per cent should be the maximum rate at which the avería should be levied.Footnote ²² The trouble was that the king had already made such promises on several occasions but had failed to keep his word. It took nearly six more years of back-and-forth negotiations with the merchants on both sides of the Atlantic before the government finally decided on the matter. Fortunately for Spain, the end of the Franco-Spanish (1635-1659) and Anglo-Spanish (1654-1660) wars cleared the way to solve the avería problem by lowering the cost of defending the fleets.

On 31 March 1660, Philip IV ordered that the registration of precious metals and other American goods would no longer be compulsory for the fleets returning to Spain. This meant that the silver would no longer be taken to the Treasure Chamber, effectively lowering the possibility of further embargoes. How, though, would the avería be levied? From then on it would be a fixed sum—790,000 ducats (1,089,154 silver pesos), the estimated cost of protecting both fleets—to be paid yearly by the Crown and the main merchant communities involved in the transatlantic trade. Each party would end up paying a different amount, although not yearly, as the Crown had intended, but only when the fleet in which they participated crossed the Atlantic (for the Andalusian merchants and the Crown this meant every fleet). In 1667, the apportionment was revised and settled for the rest of the century as follows: the merchants of Andalusia would pay 170,000 ducats (for the two fleets), the Americans 470,000 (the Peruvians alone would disburse 350,000 ducats for every fleet arriving at Portobelo) and the Crown 150,000. This agreement would be known as the asiento or indulto of the avería.

The Crown had contracted out the avería on multiple occasions, employing three-to-six-year asientos since 1591 (Céspedes del Castillo, Reference Céspedes del Castillo1945, pp. 598-601; Hierro Anibarro, Reference Hierro Anibarro2005). These contracts included the administration of everything to do with the defence of the fleets, from buying and arming warships to feeding and paying all the mariners and soldiers aboard the two flagships (known as the capitana and almiranta) and the eight-galleon armada that protected the merchant-ship convoys. A consortium of the most important merchants of Seville (with limited liability for all partners) had usually taken charge of the pre-1660 asientos, but their performance had obviously been abysmal since the 1620s. The last of the traditional asientos, that of 1640, went bankrupt as fewer and fewer merchants paid the avería. In stark contrast, the 1660 arrangement was entirely different. The measure sought to give the consulados greater control over the commercial system while leaving the administration of the defence completely in the hands of the House of Trade. The Andalusian and Spanish American consulados would be in charge of collecting their avería quotas from among the merchants of their respective community participating in each particular fleet. This was done on American soil through individual allotment (repartimiento) and in total secrecy, away from the Crown's scrutiny.Footnote ²³

The historiography has seen the 1660 arrangement, particularly the suppression of compulsory silver registration, as the Crown's final capitulation to the merchants' clandestine manoeuvres. According to Oliva Melgar (Reference Oliva Melgar2004, p. 63), it constituted «the definitive legalisation of fraud». There is no doubt that a considerable part of what had until then constituted fraud ceased to be illegal. In truth, the loosening of official control over the silver shipments—because this is what the 1660 reform meant—was an inevitable institutional adjustment following the excesses committed by the Crown since 1620. In other words, the measure made the Crown's commitment to respect property rights credible. The Council of the Indies made this point in a summary report (consulta) sent to the king in August 1661:

The measure was perhaps not a fiscal masterstroke but had clear positives for the Crown. The avería collapse of the 1640s and 1650s had forced the Royal Treasury to assume most of the cost of protecting the commercial fleets for the best part of two decades, effectively subsidising merchants (Álvarez Nogal, Reference Álvarez Nogal1997, p. 50). Without question, the 1660 arrangement was a remarkable improvement, even though the Peruvians would seldom meet their vast quota in full (Suárez, Reference Suárez2001, pp. 310-313). The Crown had such high hopes for the arrangement that 40,000 pesos of the new avería were to be subtracted to pay the juristas of the almojarifazgo de Indias—a mere 7.9 per cent of what the juristas were owed every year. It was a poor attempt to partly alleviate the fact that «no juros are paid, and [the revenue] is hardly enough to cover the levy's administration expenses».Footnote ²⁵ As far as I know, the juristas (some of them ecclesiastical institutions, no doubt) were paid only once with the avería money: 13,334 pesos brought by the New Spain fleet in October 1674.Footnote ²⁶

Crucially, there was another feature of the new avería that has gone totally unnoticed. There is widespread agreement among historians that the 1660 measure initiated the recovery of the transatlantic trade, even though the total tonnage crossing the Atlantic continued to gradually decrease (García Fuentes, Reference García Fuentes1979; Oliva Melgar, Reference Oliva Melgar2004). However, the traditional interpretation that the merchants took full advantage of a weakened monarchy and systematically got away with fraud does not explain how the extraordinary contributions (i.e. indultos, servicios and donativos), which the Crown had been able to impose using the threat of the Treasure Chamber, continued to provide a significant revenue for the Royal Treasury after 1660. In fact, I contend that fraud decreased and the Crown retained considerable bargaining power in the latter decades of the century. Although the exact overall figure has not been established yet,Footnote ²⁷ the extraordinary contributions seem to have compensated for the collapse of the almojarifazgo de Indias—although it was the Crown, not the juristas, that benefited from it, and to a considerable degree. Díaz Blanco (Reference Díaz Blanco2012, pp. 268-269) states that, in the 1680s and 1690s, there was «a very important reactivation of the tax burden on colonial trade, even to the point of not falling too far short of the intensity that was applied during the government of the Count-Duke of Olivares». The frequency and size of the payments (2,500,000 pesos in 1692, e.g.) speak volumes about the good health of the colonial trade after 1660. Moreover, the money from the Consulado became critical for the Crown, as the increasing administration and military expenses in the colonies meant less bullion was consigned to the Royal Treasury in Spain in the second half of the century (Díaz Blanco, Reference Díaz Blanco2013; Andújar Castillo, Reference Andújar Castillo2020). The contributions were usually agreed upon and channelled through a rhetoric of fraud and contraband, even though there was no surge in such infractions after 1660. The indulto (pardon) for shipping merchandise previously imported from countries at war with Spain to America became the most important type of contribution. It could not have been any other way, since barely 4.5 per cent of the total value of cargoes exported to the colonies was Spanish, and Spain was at war with France and England for considerable periods (Yun-Casalilla, Reference Yun-Casalilla1998, p. 140). Indeed, it was convenient for the Crown to use the language of fraud and contraband because it justified the indultos. How, though, did the merchants collect the money? What method did they use? Here is where the new avería came into play.

The repartimiento of the avería, which was carried out on American soil by the deputies of the Seville Consulado, served to distribute the payment of the duty individually. It was not estimated according to the silver that each person intended to ship to Spain but based on the goods they had brought to Spanish America (the deputies' assessment of the goods was different to that made by the customs officials). Individual payments were made in silver just before the fleet sailed back to Spain. The indultos, which could vary greatly in size, were added to the repartimiento. Sometimes the Consulado would borrow the money in Spain to pay the king before a fleet had even set sail for the Indies. Afterwards, the money plus the interest to pay back the lenders was raised in the main American ports using the avería's repartimiento. Two important conclusions can be drawn from this. First, contrary to what most historians have repeated, the avería continued to be an ad valorem duty after 1660. Had it not been, the individual allotments would have been impossible to devise. Second, the Consulado's increased control over the Carrera did not generate fiscal disorder but rather the opposite. The historiography insists that the 1660 arrangement signalled the Crown's weakness and that merchants took full advantage of this (Oliva Melgar, Reference Oliva Melgar, Butel and Lavallé1996, Reference Oliva Melgar2004). However, the Royal Treasury was hardly a helpless victim. If anything, the extraordinary contributions suggest that, especially after 1680, the Consulado (i.e. its board of directors) became an efficient tax collector at the Crown's service. The 1695 New Spain fleet provides a clear example: I have studied elsewhere how the deputies of the Consulado managed to collect approximately 1,800,000 pesos through a controversial repartimiento that split the merchant community into two parts (Lamikiz, Reference Lamikiz and Lanza García2023).

What happened with the almojarifazgo de Indias after 1660? Unsurprisingly, more measures were taken to undermine it and leave more room for extraordinary contributions. The 1660 decree that ended compulsory registration for the return leg also brought the abolition of the almojarifazgo de Indias for the fleets upon arrival in Spain.Footnote ²⁸ A year later, the customs value per arroba of bales and other packages containing textile products to be exported to the colonies, which had been established in 1629 but implemented from 1635 onwards, was reduced by 33 per cent (Alloza Aparicio, Reference Alloza Aparicio2019, p. 189). What is more, the customs appraisal by weight was replaced in 1680 with a method that assessed the value of the goods by their volume: 23,392 maravedís per 100 cubic palmos for goods destined for Veracruz and 36,708 maravedís if the destination was Cartagena de Indias or Portobelo. The tax farmers' complaints confirm that this measure reduced fiscal pressure on the customs houses even further.Footnote ²⁹

Is it possible to gauge the century-long effect of the reduced fiscal pressure on the customs authorities? Table 1 provides information on several bales and crates containing textile products shipped to Spanish America by four merchants (all acting as agents) in 1626, 1637, 1684 and 1699. This information has only surfaced very recently in notarial and private archives. Previously, historians knew virtually nothing about the nature and market value of the merchandise exported to the colonies in the 17^th century. Although it is a small sample and more findings will no doubt help to refine my interpretation, Table 1 shows that the customs value of goods shrank massively in the second half of the century. The merchandise that Domingo de Olea brought to the customs house in 1626 was declared at slightly over 40 per cent of the invoice prices. The method of appraisal by weight had not been introduced yet, so the customs officials—since they could not inspect the goods—had no choice but to rely on both Olea's word and the packages' appearance. It was a flagrant example of tax evasion. By 1637, valuation by weight had been recently introduced. However, Gregorio de Ibarra's bales and crates were also valued at slightly over 40 per cent of the invoice prices. The contrast with the merchandise shipped in 1684 and 1699 is stark. The bales and crates that Bernardo Clemente Príncipe and Juan de Iriarte took to the customs house were valued at between 2.3 and 5.7 per cent of their invoice prices. At first glance, this looks like outrageous tax evasion, a view to which most historians of the Carrera would traditionally subscribe without reservation. In fact, however, there was no fraud in the 1684 and 1699 customs values, which were instead the product of pricing goods by their volume, in accordance with the method introduced in 1680. In other words, Príncipe and Iriarte paid what was due.

Table 1. Invoice and customs values of several bales and crates containing textile products exported to Spanish America in 1626, 1637, 1684 and 1699 (In Silver Pesos of 272 Maravedís)

Sources: Díaz Blanco (Reference Díaz Blanco2022); Díaz Blanco and Hernández Rodríguez (Reference Díaz Blanco and Hernández Rodríguez2022); Lamikiz (Reference Lamikiz and Lanza García2023); and «Quenta y ajustamiento de … Gregorio de Ybarra», Archivo General de la Nación-Perú, Consulado 208-8, ff. 204r-209v. Note: 1 tonelada de arqueo of 1.376 m³ was equal to 166.375 cubic palmos.

Although the sample of bales and crates contained in Table 1 is small, another conspicuous feature deserves consideration. Does the sample show that more expensive textiles were shipped to Spanish America as the century progressed? Unfortunately, comparing the prices of the bales' contents is problematic because the names of the textiles differ. However, there are two alternative ways to compare the older and newer merchandise. First, by juxtaposing the invoice values per unit of volume. In the case of the 1637 goods (1,095 cubic palmos of ropas, worth 11,959 pesos), their value per cubic palmo was 10.9 silver pesos. In contrast, the combined average value per cubic palmo of the bales shipped in 1684 and 1699 (557 cubic palmos of ropas, worth 16,152 pesos) was 29 silver pesos. In other words, the invoice value per unit of volume had almost tripled in about five decades (while the fleets had lost as much tonnage in the same period). However, the second way to compare the merchandise suggests that, in the sample, the value increase per volume was caused by the 1680 customs appraisal method. I have the names and prices of all the textile items contained in the 1637 and 1699 bales. Textiles measured in varas (0.835 metres per vara) made up 86 per cent of the value of the former, and 92 per cent of the latter. What was the average price per vara in each case? Eight silver reales in the 1637 bales and 6.41 in the 1699 ones. This means the 1637 vara textiles were more expensive than those of 52 years later. How is this possible if the value per volume had increased? The explanation is that the 1699 bales had many more textile varas packed in the same space, the natural consequence of taxing goods per volume instead of by weight. It could also have been because the width of the textiles narrowed over time. This demands more research.

The small size of the sample demands extreme caution when drawing any conclusions. The sample suggests that ordinary fiscal pressure was gradually and officially lowered in the 17^th century while, at the same time, a transformation in the value of goods per unit of volume was also taking place.