Lending to the Borrower from Hell: Debt, Taxes, and Default in the Age of Philip II (The Princeton Economic History of the Western World)

by Mauricio Drelichman

  By their very nature, asientos were riskier than juros. Although the contracts might specify the intended sources of repayment, no funds were specifically earmarked on signing. If the cash flow situation was critical, the treasurer might delay issuing the repayment orders. Even after excluding contracts affected by the defaults, over 20 percent of asientos issued during Philip II’s reign were not repaid by the deadline originally specified. Payments could be delayed from a few weeks to several years. In many cases, the king and bankers would negotiate a consolidation of outstanding payments into a new asiento, with additional interest added to compensate for earlier, unmet obligations.

  A NEW SERIES OF ASIENTOS

  In terms of data, an important contribution of this book is a new series of asientos, consisting of every contract between Philip II and his bankers between 1566 and 1598. Two additional years from the reign of Philip III, 1599 and 1600, were added to allow for a better exploration of the effects of the 1596 bankruptcy.

  Studying asientos is vital to understand the financing of the early modern Castilian state as well as the workings of the early Spanish Empire. While asientos represent a minority share in total debt, ranging between 10 and 25 percent of all obligations, they are the segment that exhibits by far the greater volatility. Asientos are where the action is. They also experience a wide array of rescheduling and recontracting. While juros were pretty much honored on time and in full throughout the sixteenth century, the letter of asientos was not observed to some degree almost 30 percent of the time. Their continued popularity in the face of this volatility as well as their apparent insecurity seems at odds with the conventional wisdom of modern international finance, calling for a deeper investigation. Finally, knowing the aggregate value of asientos issued each year is an essential step in determining the fiscal position of early modern Castile. In a context of scant aggregate macroeconomic indicators, our series serves as the linchpin that allows us to perform the earliest reconstruction of full yearly national accounts for any sovereign state in history.

  Asientos were issued in duplicate. The original document, signed by the king, was given to the bankers who provided the funds, while a copy was kept in the Contaduría Mayor de Cuentas—the body in charge of auditing all the financial transactions of the Crown. These copies have been preserved in the Archive of Simancas, where a relatively recent cataloging effort has produced a complete series of asientos beginning in 1566.42 We collected every asiento in this series up to the year 1600—amounting to 438 documents and over 5,000 manuscript pages—and transcribed their every single clause. We recorded up to 90 variables for each contract, including the date, the bankers’ names, the amounts involved, the delivery and repayment locations, the tax streams that were supposed to provide the repayment, the type and amount of collateral (if any), whether the contract involved an exchange operation, the presence and nature of contingency clauses, and a variety of other descriptors. We then used the text in the clauses to reconstruct the agreed-on cash flow for each contract, which will help us establish their actual rates of return.43

  Ours is not the first effort to describe short-term borrowing during Philip II’s reign. Ulloa (1977) compiled a series of asientos between 1566 and 1596, relying on the same documents we used, but his work suffers from serious shortcomings. First, Ulloa overestimates the total amount of short-term borrowing as a result of including asientos entered into by field commanders in Flanders with the agents of Genoese or Spanish banking houses. These loans were typically consolidated into larger contracts in Madrid, and hence are counted twice in Ulloa’s series.44 Next, Ulloa relied only on the summary information given in the first page of the contracts. Asientos were used for a variety of purposes, including lending, transferring money, farming out tax revenues, and chartering monopolies. A single contract would often involve two or more of these transaction types; the lending and transferring combination was particularly widespread. The summary information on the first page lists the total monetary amounts involved without distinguishing among transactions. By relying on these headline figures, Ulloa overestimated the total amounts lent. By actually reading the contractual clauses, we are able to separate borrowed amounts from all the other transactions involved. Finally, Ulloa did not collect any information on terms and conditions, interest rates, or other variables that might shed light on the workings of asientos.

  Earlier scholars have already compiled time series of asientos outstanding for the period before Philip’s reign. Carande (1987) pioneered the study of time series with his work on the loans of Charles V. Charles’s documents are still today not organized into a series; collecting them was a painstaking multiyear effort, in which it was necessary to track down every individual contract. While it is unfortunate that Carande’s study does not include enough information to calculate the interest rate on the contracts, his work nevertheless remains a standard source on Castilian short-term borrowing between 1519 and 1555.45 The archival record between 1555 and 1566 is incomplete, and has not been systematically cataloged; estimating the amount of short-term debt subscribed during this period is difficult.46 Carlos Alvarez Nogal (1997), Sanz Ayán (1998), and Juan Gelabert (1999b) have compiled series on the short-term borrowing of Philip III and Philip IV.

  Figure 6 shows the annual real value of asiento issues between 1566 and 1600, separating their gross value from their loan component.47 The first striking feature is the lack of a trend—Philip II borrowed as much per year at the end of his reign as he did at the beginning. Next, the bankruptcies of 1575 and 1596 only result in short interruptions in lending. New issues resume in 1577 and 1597, respectively. The high borrowing volumes in these last two years reflect the new loans negotiated with bankers as part of the respective settlements.

  FIGURE 6. Yearly asiento issues, 1566–1600

  Annual short-term borrowing decreased substantially between 1577 and 1584, and again in 1597–99, relative to the previous trends. Did access to credit suffer after the defaults? This is unlikely, for two reasons. For one, with each settlement Philip received a fresh loan, worth 5 million current ducats in 1577 and 7.2 million in 1597. These amounts, captured in the respective columns of figure 6, are comparable to the peak volume of predefault lending. Furthermore, both ordinary tax revenues and silver remittances were unusually strong in the years 1576–81 (figure 4). Revenue from sales taxes grew from 1.1 million ducats in 1575 to 3.2 million in 1576 and 1577 before settling down to an annual rate of 2.4 million—more than twice its predefault level. Silver revenue also surged, amounting to almost 2 million ducats in 1577 (compared with an average of 0.7 million between 1570 and 1575). Overall, lending during the eight years after 1576 declined by 2.1 million ducats annually compared to the prior period. Annual revenue was up by 1.8 million ducats. In addition, warfare in the Low Countries declined following the Pacification of Ghent. Lower borrowing does not imply that the Crown was shut out of credit markets. Rather, the elimination of the deficit through a combination of revenue windfalls and lower expenditure made further borrowing unnecessary.

  Asiento issues reflected military events. In 1567, the start of the Dutch Revolt drove up short-term borrowing to its highest value in the entire period. Issuance remained high throughout the early 1570s, with Castile fighting in both the Low Countries and Mediterranean. The temporary withdrawal from Flanders after the Pacification of Ghent resulted in a large “peace dividend”: borrowing was much lower in 1576–84. The resumption of hostilities in Flanders along with the outfitting and rebuilding of the Armada increased borrowing once again to high levels in 1585–91. Renewed fighting against England and France spelled sustained high asiento issuance until Philip II’s death and beyond.

  CONCLUSION

  When Philip II acceded to the throne in 1566, Castile was already a major player on the European political and military scene. The reforms of the Catholic Kings in the later part of the fifteenth century were instrumental in increasing state capacity early on. The military campaigns of Char
les V, though expensive and of mixed military value, put European powers on notice of Castile’s power and imperial ambitions. These were backed by a fast-growing economy that, along with a variety of fiscal instruments, allowed a comparatively nimble bureaucratic machinery to increase revenues and use them to project military might. The state itself was built on a broad support base, with the Crown keeping a large but not unchecked role. Through their participation in the Cortes, mercantile interests could and did limit the king’s ability to tax. Fiscal negotiations were meaningful. Whenever the reason for additional revenues was not convincing, the Cortes used its effective veto power. At critical junctures, when the survival of the state was at stake, additional revenues were granted in exchange for administrative and supervisory powers as well as other concessions. The Cortes also placed a limit on the revenues that could be used to guarantee the payment of juros, effectively setting a ceiling for long-term debt. These features provide little support for the literature’s characterization of Castile as an unqualified absolutist state.48

  The Castilian portfolio of fiscal and financial instruments was remarkably complete for its time. The revenue technology was a mix of useful adaptations of medieval legacy taxes combined with newer excises and trade duties. As the economy grew, the focus of taxation shifted increasingly to internal and overseas trade, with American silver eventually providing up to 25 percent of all fiscal revenues. On the financing side, the Crown had both long-and short-term debt instruments at its disposal. Although the long-term juros were technically nominative, they were widely traded in the secondary market, supplying a key element for the correct pricing of debt. Since they were held by Castilian elites and backed by earmarked revenues, they were perceived to be safe. This perception proved correct, as juros were strictly honored throughout the sixteenth century. On the short-term front, asientos allowed the Crown to smooth the volatile silver revenues, leverage income outside the purview of the Cortes, and quickly shift resources throughout the empire. The contractual complexity of these loans, allowing for a variety of contingent delivery and repayment scenarios while spreading the risk among multiple layers of lenders, rivals that of sophisticated modern instruments.

  Castilian asientos did not become famous because of their innovative contractual structures. They instead earned their celebrity status through their central role in the seven bankruptcies declared by the Crown between 1557 and 1647. In the chapters that follow, we reconstruct their anatomy, analyze their causes and consequences, and shed light on the role they played in the international financial architecture built around the most powerful empire of its time.

  1 Acemoglu, Johnson, and Robinson (2005), for example, code Spain as being an “absolutist” state.

  2 See, for example, the definition of absolutism by Richard Bonney (1987).

  3 We cannot review the entire literature on the reality of absolutism here. For good overviews, see Oestreich 1969; Parker 1983.

  4 We provide a full set of international comparisons in chapter 8.

  5 Technically all executive decisions emanated from the king, while councils were limited to the consultative function implied in their name (consilium). The actions of the councils were called consultas, which implied a request for or offer of advice. In practice, the sheer volume of state business meant that all but the most crucial decisions were taken by the councils and rubberstamped by the king.

  6 In our description of the system of councils, we follow Artola 1988.

  7 Certain revenue streams were administered by bodies other than the Council of Finance. This was the case of revenues collected through the church, which were overseen by the Council of Crusade; later on, the excises known as the millones would be overseen by the Diputación, an organ of the Cortes.

  8 The Cortes has been the subject of a large body of scholarly research, with novel contributions appearing to this day. For some general treatments centered on the early modern period, see Carretero Zamora 1988; Thompson 1976, 1993, 1994a; Fortea Pérez 2009; proceedings of the Congreso Científico sobre la Historia de las Cortes de Castilla y León (1989). The first documented meeting of the Cortes took place in León in 1188.

  9 The Cortes had to be called by the king and could not convene of its own accord. A number of taxes would technically lapse, though, if not renewed by the Cortes, and hence monarchs had powerful incentives to call regular meetings of the assembly.

  10 See Jago 1985; Fortea Pérez 2009.

  11 These events support David Stasavage’s (2011) conjecture that geographic distance could be a major obstacle for a representative assembly trying to monitor and constrain a monarch in the early modern period.

  12 For an overview of early modern Castilian nobility, see Domínguez Ortiz 1985.

  13 The sale of privileges of lordships over formerly free towns caused serious distress for local economies, as towns lost income streams to their new overlords or saw their fiscal pressure increase. Several towns enacted extraordinary levies in an effort to buy their own privileges and remain free from external overlordship. For a discussion on the sale of towns under the Habsburgs, see Nader 1990.

  14 For more on the privatization of tierras baldías, see Vassberg 1975, 1984; García Sanz 1980.

  15 For an analysis of the expansion of the number of hidalgos in the sixteenth century, see Drelichman 2007.

  16 The share of clerics was estimated at 1.5 percent of the population. The number rings small, possibly because ecclesiastics succeeded in underreporting their numbers and hence lowering their contribution to the new tax. The figure is still an order of magnitude lower than the percentage of land controlled by the church.

  17 For the figures for Seville, see Benassar 2001. For Toledo, see Drelichman and González Agudo 2013.

  18 Jordi Vidal-Robert (2011) provides an economic analysis of the Spanish Inquisition as a tool of social control.

  19 For a modern estimate of the victims of the Inquisition and composition of the sentences handed out, see Pérez 2003.

  20 The discussion of revenue sources is based on Thompson 1976; Ulloa 1977; Artola 1982, 1988. Quantitative data are drawn from our previous work in Drelichman and Voth 2010.

  21 Originally the excusado was to be equal to the tithe of the richest parishioner. In a reflection of the extreme skewness of the income distribution, the bishops fought hard to have it changed to the tithe of the second-richest parishioner. Since the Crown and bishops negotiated a lump-sum payment in lieu of the tax, the point was moot in regard to Crown revenue. It nonetheless remained important for local excusados, such as those used to finance the fabricae of large cathedrals.

  22 Migratory sheepherding entered a declining phase in the second half of the sixteenth century and dropped precipitously after 1580. The activity had long ceased to be a source of significant Crown revenue by then (Drelichman 2009). The 1568 rebellion of the Morisco population in the Alpujarras region resulted in both the diaspora of the Moriscos and destruction of mulberry trees, fatally damaging both the silk industry and the taxes it generated.

  23 The mercury amalgam smelting process permitted the cost-efficient refining of low-grade ores. It was described in Agricola’s De Re Metallica, published in 1556, and it seems certain that Medina had access to this knowledge before setting out for Mexico in 1554. The chemistry of the patio process has only recently been replicated in a laboratory setting and thoroughly dissected. For a technical treatment, see Johnson and Whittle 1999.

  24 Earl Hamilton’s classic account of silver imports into Castile relied on official records from the House of Trade. While these records do not reflect contraband, they capture royal silver income accurately. See Hamilton 1934; Ulloa 1977.

  25 See the discussion in chapter 2. Morineau (1985) does not find significant discrepancies between declared amounts and total shipments reconstructed from Dutch gazettes until the 1580s, leading to the conclusion that smuggling was limited up to that time.

  26 For the full reconstruction, including yearly data, se
e Drelichman and Voth 2010. For individual revenue stream data, see Thompson 1976; Ulloa 1977. The deflator is the price index for Old Castile reported in Drelichman 2005.

  27 The Cortes of 1574 was studied by A. W. Lovett (1980). Charles Jago (1985) and José Ignacio Fortea Pérez (2009) provide accounts of the Cortes of 1576.

  28 For an overview of juros and their history, see Toboso Sánchez 1987. The practice of granting rights to future fiscal revenue in exchange for an upfront payment seems to have first originated in twelfth-century Genoa, although the final characteristics of juros are observed evolving simultaneously in several European locations.

  29 Lifetime juros could only be transferred in special circumstances, as their value depended on the age of the holder.

  30 The agreement is detailed in a contract between the king and Jerónimo de Salamanca, Lucas Justiniano, and Bautista Spinola dated May 19, 1569. AGS Contadurías Generales, Legajo 85.

  31 There is no central registry of juros. The cataloging does not allow us to identify bonds belonging to a particular period; this can only be done poring by hand through each document. Furthermore, the available juros are not a random sample; only bonds that were resold at some point in time are included. A bond that was resold many times would register several duplicate instances in the archival record, with little information to link them together. On the other hand, ecclesiastical institutions and noble families were large investors, and many were unlikely to ever trade the bonds they held, and hence their juros would be missing from any attempt at a reconstruction.

  32 For the standard source on the asientos of Charles V, see Carande 1987.

  33 This is not to say that Charles V serviced every single loan at all times; for example, the Fugger were still waiting for repayment of debts from the Schmalkaldian War in 1551 some seven years later (Ehrenberg 1896).