Lending to the Borrower from Hell: Debt, Taxes, and Default in the Age of Philip II (The Princeton Economic History of the Western World)
Page 5
States run by a successful military-fiscal complex eventually dominated the map of Europe. What had been a patchwork of different territories consolidated into larger territorial units. For example, England and Scotland became unified after the Act of Union in 1707, and Poland disappeared from the European map altogether. Consolidation in the nineteenth century became even more rapid, as both Germany and Italy turned into nation-states. This process of consolidation interacted with centralization and military developments. States that lagged in developing an effective fiscal system and competent army were eventually absorbed by more successful powers.
The costs of war were not limited to the pressures they put on the public purse. War could cause massive population losses and severe economic disruption. Borrowing diverted funds from civilian to military uses. In Britain, “crowding out”—the substitution of public expenditure for private investment—likely occurred (Williamson 1987; Temin and Voth 2005). The medium- and long-term effects of war were more limited. In an agrarian society, population losses had a silver lining, raising per capita incomes: fewer people cultivated the same amount of land (Clark 2007; Voigtländer and Voth 2013). Wooden houses could be rebuilt quickly. Overall, there is agreement in the literature that the effects of war were not overwhelmingly negative for output. In the words of Jan De Vries (1976), “It is hard to prove that military action checked the growth of the European economy’s aggregate output.”65
Some scholars have even argued that the ruthless pursuit of military power—extended to trade relationships in the eighteenth century by mercantilists—also laid the basis for greater economic riches. Military technology advanced faster and further in Europe after 1500 than in the rest of the world—probably as a result of a highly belligerent environment (Hoffman 2011). There may have been technological spillovers to private industry.66
In the case of early modern city-states, Ehrenberg (1896, 6) argued that “credit was [the] strongest weapon in the fight for freedom.” Princes similarly depended on credit to outspend and outfight their rivals. Without the ability to increase spending rapidly in wartime, military defeat loomed large—followed by losses of reputation, treasure, and territory. For each political entity, access to borrowing was a godsend. As states and cities borrowed more, however, they drove up the length, intensity, and cost of war. Fiscal exhaustion as much as defeat on the battlefield sealed the fate of the weaker contenders; as a Spanish military commander eloquently put it, “Victory will go to whoever possesses the last escudo” (Parker 1998).67 It was against this backdrop that Philip II ran up enormous debts and defaulted on them four times, while continuing to expand what was already the largest empire on earth.
THE SPANISH CASE: THE DEBTS AND DEFAULTS OF PHILIP II COMPARED
Our book is a case study of a single sovereign borrower—Philip II of Spain. How does his borrowing history compare with the general pattern of cross-border debt and default between 1800 and the present? Table 2 presents summary statistics for the key indicators of sovereign debt in the past two centuries, side by side with the same figures for Philip II’s reign.
The reign of Philip II registered frequent defaults; more than one year out of five saw some kind of payment suspension. And yet the tumultuous history of sovereign debt since 1800 registers similar ratios. In the data set used by Reinhart and Rogoff, roughly 20 percent of countries were in default on average in any given year. In general, countries with a history of defaults suspended payments again (Lindert and Morton 1989; Reinhart, Rogoff, and Savastano 2003); in the case of Philip II, there was a sequence of four defaults. These are explored in more detail in chapter 4, and constitute the start of what went on to become a record-setting sequence of thirteen suspensions: no country in recorded history has defaulted more times.
Table 2. Philip II and the general pattern of debt and default since 1800
Issue
General pattern
Philip II
Frequency of default
20%a
21.4%i
Serial default probability
35–80%b
100%
Haircuts
6.9%–72.9%c
20–38%j
Average length of restructuring period
8 yearsd
2.25 years
Interest rate penalty after default
None
None
Average debt-to-GDP ratio
52.8%e
~60%k
Fiscal revenue-to-GDP ratio
16.3%f
~10%k
Rate of return (nominal and promised ex ante)
4.9–6.9%g
10.3–24.1%l
Excess return (real and ex post)
-1.92–2.53h
3.16%
a Reinhart and Rogoff 2009. Michael Tomz and Mark Wright (2007) calculate that there were 250 defaults in a set of 175 countries since 1820. This implies a much lower default rate; if the average duration of a default was eight years, then there would have been a default in only 6.3 percent of the country years. The reasons for the discrepancy are most likely that many of the 175 countries in their database are relatively young and that the true number of potential default years is much lower than implied in our calculation.
b Tomz and Wright 2007. Calculated as the frequency of default, given an earlier default episode.
c Sturzenegger and Zettelmayer 2008.
d Benjamin and Wright 2009.
e World Development Indicators. Includes countries with information on central government balance as well as general government debt.
f World Development Indicators. Average for all countries with observations in 1960–2000.
g Lindert and Morton 1989.
h Lindert and Morton 1989.
i Philip II reigned for forty-two years and was in default during nine of them (across four bankruptcies).
j Third and fourth defaults only.
k GDP figures are highly uncertain; we use estimates in Alvarez Nogal and Prados de la Escosura 2007.
l Twenty-fifth to seventy-fifth percentile of the modified internal rate of return (MIRR).
Throughout history, the size of haircuts has varied by at least a factor of ten—from less than 7 percent to more than 70 percent. Philip II’s defaults were relatively mild by modern standards; the biggest and most severe rescheduling, the medio general of 1575, led to average losses of 38 percent. The 1596 default was even milder and resulted in losses of 20 percent in net present value (NPV) terms.
Restructurings under Philip II were also surprisingly swift. The average renegotiation in the recent past took eight years (Tomz and Wright 2007).68 Philip II and his bankers typically settled within twenty-six months—a length of time that declined with each default. The 1596 episode was resolved within a year. Interestingly, countries that defaulted in the past typically did not experience an increase in the interest rates at which they could borrow after a resumption of lending (Lindert and Morton 1989). This was also true in the case of Philip II, where lending conditions remained broadly unchanged before and after the well-documented defaults of 1575 and 1596.69
Debt-to-GDP ratios have fluctuated greatly in the past. On average, countries have been able to borrow just over 50 percent of their output. Philip II accumulated substantial debts by the standards of the age. With the total value of outstanding debt close to 60 percent of GDP, Habsburg Spain could have qualified for initial euro membership, but only just. For an early modern state, accumulating such debts should perhaps be thought of as a sign of success—of an administrative machine producing fiscal revenues on a scale large enough to make repayments a distinct possibility in the eyes of lenders (Besley and Persson 2010). Certainly the ability of the Spanish monarchy to raise and allocate around 10 percent of GDP in taxes and contributions was unusual at the time, and must be recognized as a major accomplishment.
On average since the nineteenth century, lending to sovereigns has been profitable, but this generally positive exp
erience has been punctuated by periods of sharp losses. The same was true of lending to Philip II. We calculate that bankers on average made greater profits than what was available elsewhere (chapter 6). These were also slightly higher than those found in the more recent past.70 While Peter Lindert and Peter Morton (1989) document excess returns of up to 2.5 percent a year, we find an annual value of 3.16 percent over a thirty-year period.71
To sum up, the borrowings of Philip II of Spain seem remarkably familiar to the modern eye. The debt levels, size of haircuts, rates of return, interest rate penalties after payment suspensions, and default frequencies were not fundamentally different from what scholars have documented for the two centuries after 1800. That so many outcomes are similar suggests that the forces at work today are the same ones that allowed the market for sovereign debt to thrive in sixteenth-century Habsburg Spain.
THE FOUR QUESTIONS
This book’s central concern—why Philip II could borrow so much while defaulting so often—is broad and complex. We break it down into smaller, more digestible issues and structure the book around four key questions. Did Philip II have enough resources to pay back his debts? If yes, what prevented him from behaving opportunistically? (That is, why did he not just default to make himself better off?) What were the benefits for the bankers? How can we interpret the recurring defaults? In the language of economics, these questions address the sustainability of debt, the compatibility of incentives, the profitability of lending, and the nature of contracting.
COULD PHILIP II PAY BACK HIS DEBTS?
Our first order of business is to establish whether the Crown of Castile had the capacity to actually service and repay its debts. If bankers lent to a monarch who could not have possibly repaid his debts, then bankruptcies are a foregone conclusion. Financiers would have to be irrational (in an economic sense) to extend credit to such a sovereign, but this would hardly be the first—or last—time that lenders did not act in their own best interests. Indeed, repeated defaults have been invoked as a sign of irrationality by many historians and economists.72
In macroeconomics, the term “sustainability” describes a situation when debts can be serviced indefinitely into the future. We therefore ask whether Castile’s debt load was sustainable, using the standard tools and tests of modern international finance. To that end, we construct a set of national accounts for Castile at an annual frequency between 1566 and 1596. This thirty-one-year data series is the earliest of its kind for any sovereign state in history completed so far.
Armed with our fiscal data, we show that Castile passes several tests of debt sustainability, and that these results resist various robustness checks. In many ways, the results of our accounting exercises are too pessimistic to judge the actions of sixteenth-century actors. Our sustainability standard is the same one applied by the International Monetary Fund (IMF) when evaluating the fiscal solvency of modern states. It is both stringent as well as ahistorical; Philip II’s finances pass this test with flying colors.
Castile in the second half of the sixteenth century raised much more money in taxes than it spent on the armies, royal court, and law enforcement. In other words, it ran massive primary surpluses, which could be—and were—used for debt service. On average, Castile’s growing debts between 1550 and 1600 were well supported by rising tax revenues; as the debt burden grew, the country generated higher and higher primary surpluses. Both tax increases and silver revenues from the Americas helped to keep the weight of rising debt in check. The Castilian debt burden in the sixteenth century was no higher overall than that of highly successful nineteenth- and twentieth-century economies.
Castile also spent much more than could be expected on its wars, while receiving almost no monetary returns from them. Had some campaigns been more successful or just shorter, the decline in expenditure would have vastly improved the Crown’s already-adequate ability to service its loans. Lenders were not wrong in expecting that Philip II would have enough resources to pay them back in the long run. Although there were periods when liquidity was tight, on average the money to service the asientos was there. The Crown’s success in supporting functioning markets for the various types of debt it issued is testimony to its sound fiscal footing.
WHY DID PHILIP II WANT TO PAY BACK HIS DEBTS?
The Crown may have been able to service all its debts in the long run, but this doesn’t mean that it was in its best interest to do so. A sovereign is a special kind of borrower—one that cannot be taken to court or otherwise forced to honor their contracts. This can lead to opportunistic behavior. Plenty of European monarchs refused to pay back their debts, sometimes taking the lives of creditors—as well as their money—in the process. If Philip II faced the same incentives, lending to him could have been not just economically irrational but also downright foolish. When a monarch averages one default a decade, there is enough evidence to suspect opportunistic behavior.
Opportunistic defaults only work if a borrower has an outside option. The king, after confiscating the loans from a set of lenders, must either be able to turn to someone else capable of satisfying their financing needs or attempt to provide for them in-house. Neither option was available to Philip II. Lending in the sixteenth century occurred in an “anarchic” environment (Kletzer and Wright 2000); neither king nor bankers could credibly commit. In actual fact, both bankers and king defaulted on the letter of agreements; one banker went bankrupt, never returning the funds deposited by Philip. This means that the king could not simply default, save the interest, and draw down the balance whenever the need arose. If there is also no alternative lender to turn to, then maintaining a good reputation—by faithfully servicing one’s debts—is the only way to maintain access to smoothing services (Cole and Kehoe 1995).
The Genoese bankers clearly understood this and acted accordingly. The archival documents we explore show a tightly interwoven network of bankers. The incentives among lenders were carefully structured to ensure mutual support. Crucially, no member cut side deals during payment suspensions; moratoriums on lending to the king had “bite” because no other lender entered the market. This is true of Genoese network members; it is also true of outsiders, such as the Augsburg banking house of Fugger. Their correspondence suggests that a “cheat-the-cheater” mechanism (Kletzer and Wright 2000) was responsible for this—the very real fear that if they extended fresh credit while the king had suspended payments on the Genoese, they would be next in being defaulted on.73 This ensured that the incentives of the king and bankers were aligned, and that debts were repaid as soon as the resources to do so were available.74
Why was being cut off from loans so painful to Philip II? The answer, in one word, is war. All borrowing is about breaking the link between revenue and spending; to shift expenditure from one period to the next is particularly useful in times of war, when financial demands can be urgent indeed. Philip II lived in one of the most belligerent periods in all of recorded history. The ability to ramp up spending in times of military necessity was valuable; without a viable alternative, a working relationship with the Genoese bankers was essential.
James Conklin (1998) has argued that the reason why Philip II repaid his debts was that the Genoese could impose costly sanctions on him. In his interpretation, the suspension of transfers following the default of 1575 caused a sharp reversal of Spain’s military fortunes in the Netherlands; unpaid troops mutinied and sacked the loyal city of Antwerp, causing a massive backlash against Spanish rule. Afterward, the king quickly settled. We do not find support for Conklin’s interpretation. A close reading of the documentary record shows that the “transfer stop” by the Genoese had no bite, transfers continued unabated, and the king’s treasurer in the Netherlands had ample funds to pay the mutinous troops. A power vacuum caused by the death of the governor-general in Flanders instead allowed matters to get out of hand. This explains the mutinies and why the events in Antwerp had little impact on the eventual default settlement.
WHY DID BANKERS
LEND?
The king could pay, and had the incentives to do so, but any lending relationship needs two willing participants. The suspension of 1575 imposed an average loss of 38 percent on all outstanding debt—a large hit that could have discouraged lenders from trying their luck again. Yet our data show that bankers from the same families lent to Philip for the duration of his reign, regardless of how many of their members were caught in a default or how much of their capital was compromised. Why was lending to the first serial defaulter in history so attractive?
To answer this question, we analyze the asientos clause by clause, reconstructing their agreed-on cash flows. This allows us to calculate the promised rates of return, which we then modify depending on whether a contract was affected by a default. When we aggregate the loans by banking family, the puzzle is resolved. Returns during normal times were high—very high. By any standard, they more than made up for the losses incurred during the bankruptcies. A suspension could inflict severe short-term losses, but these would be amply compensated as long as the family continued to lend to the king for a sufficiently long time. By the standards of the time, it was notable that no banker lost his life lending to Philip II (the murderous thoughts of the Marquis of Poza notwithstanding).75 We find something even more astounding: despite the less than stellar repayment record of the Crown, almost no banking family lost money.
WHAT WAS THE NATURE OF THE DEFAULTS?
Castilian debts were sustainable, the king had appropriate incentives to honor them, and bankers made healthy profits on their loans. Why, then, did Philip II default four times during his reign, and how should we think about these episodes? Understanding the nature of these defaults as well as their implications for theories of sovereign finance is the ultimate goal of our book.