12 International Organization increasing the ability of those with a stake in the repayment of debt to impose penalties on him.By raising the maximum value of P,institutions provide greater inducements for the sovereign to honor his loan agreements and,thereby,render his promises to do so credible.Higher penalties raise the maximum amount of debt that the sovereign will credibly repay and hence the amount lenders will pro- vide.Institutions can thus serve as a form of commitment technology by taking discretion away from the ruler and increasing his incentive to carry out his contracts. Throughout history,a number of institutional mechanisms have been created to serve this function.Root and Hoffman,for example,argue that the French mon- archy fostered corporate lending structures that made it easier for potential credi- tors to boycott the crown in the event it reneged on their loans.43 Given that these structures were intended to operate within a monarchical system,they are clearly not"democratic"or "liberal"in the usual sense.Nevertheless,the institutions as- sociated with representative political systems are particularly well suited to serve as commitment mechanisms.44 Two related features of liberal government are particularly salient in this re- spect:the diffusion of political authority to a parliament or representative legisla- ture and the establishment of low-cost mechanisms for sanctioning representatives and state leaders.Effective legislative assemblies with some"power over the purse" place an important constraint on our hypothetical sovereign.No longer can he unilaterally decide matters of fiscal policy;instead he must bargain with a repre- sentative assembly on such issues.At a minimum,this creates a new "veto player" who can prevent actions that are contrary to constituents'interests.In more ad- vanced democracies,power over the purse can entail considerable influence in the drafting and passage of budgets. Representation is further guaranteed by the introduction of low-cost mechanisms-such as elections-for sanctioning those officeholders who renege on their agreements.To the extent that politicians value holding office,electoral institutions help to align the incentives of representatives with the interests of their constituents.In modern democratic systems,this form of sanction can be imposed on the executive as well,whether that authority resides in a parliamentary cabinet or an elected president.In these cases,the unaccountable sovereign disappears altogether,and all state officials can be penalized in a relatively easy manner. In the emerging liberal states of early modern Europe(for example,seventeenth- century Holland and eighteenth-century Britain),representation was limited to sub- stantial wealth-holders.In the case of debt,this had a remarkably salutary effect, for-as detailed below-default required not only the inclination of the crown, but also the approval of parliament.In effect,this gave the representatives of the bondholders a veto over decisions about honoring loan agreements,dramatically lowering the probability of default.45 43.Root 1989;see also Hoffman 1994. 44.North and Weingast 1989. 45.Ibid
12 International Organization increasing the ability of those with a stake in the repayment of debt to impose penalties on him. By raising the maximum value of P, institutions provide greater inducements for the sovereign to honor his loan agreements and, thereby, render his promises to do so credible. Higher penalties raise the maximum amount of debt that the sovereign will credibly repay and hence the amount lenders will provide. Institutions can thus serve as a form of commitment technology by taking discretion away from the ruler and increasing his incentive to carry out his contracts. Throughout history, a number of institutional mechanisms have been created to serve this function. Root and Hoffman, for example, argue that the French monarchy fostered corporate lending structures that made it easier for potential creditors to boycott the crown in the event it reneged on their loans.43 Given that these structures were intended to operate within a monarchical system, they are clearly not "democratic" or "liberal" in the usual sense. Nevertheless, the institutions associated with representative political systems are particularly well suited to serve as commitment mechanism^.^" Two related features of liberal government are particularly salient in this respect: the diffusion of political authority to a parliament or representative legislature and the establishment of low-cost mechanisms for sanctioning representatives and state leaders. Effective legislative assemblies with some "power over the purse" place an important constraint on our hypothetical sovereign. No longer can he unilaterally decide matters of fiscal policy; instead he must bargain with a representative assembly on such issues. At a minimum, this creates a new "veto player" who can prevent actions that are contrary to constituents' interests. In more advanced democracies, power over the purse can entail considerable influence in the drafting and passage of budgets. Representation is further guaranteed by the introduction of low-cost mechanisms-such as elections-for sanctioning those officeholders who renege on their agreements. To the extent that politicians value holding office, electoral institutions help to align the incentives of representatives with the interests of their constituents. In modern democratic systems, this form of sanction can be imposed on the executive as well, whether that authority resides in a parliamentary cabinet or an elected president. In these cases, the unaccountable sovereign disappears altogether, and all state officials can be penalized in a relatively easy manner. In the emerging liberal states of early modern Europe (for example, seventeenthcentury Holland and eighteenth-century Britain), representation was limited to substantial wealth-holders. In the case of debt, this had a remarkably salutary effect, for-as detailed below-default required not only the inclination of the crown, but also the approval of parliament. In effect, this gave the representatives of the bondholders a veto over decisions about honoring loan agreements, dramatically lowering the probability of default?' 43. Root 1989; see also Hoffman 1994. 44. North and Weingast 1989. 45. Ibid
Institutional Foundations of Financial Power 13 In modern democracies,a much wider range of interests is represented.There is nothing inherent in such systems to prevent nondebtholders from seeking to de- fault,especially when the debtholders are not an important constituency of the party in power.Nonetheless,default is costly to the rest of the economy and hence to other important constituencies such as labor or nondebtholding capitalists.46 Economists list several such costs,and two are relevant for us.47 First,default implies that debt finance will be more difficult in the future.Any future incre- ments to spending are therefore likely to require new taxes.Moreover,the demise of debt finance removes the state's ability for tax smoothing,implying lower eco- nomic growth and hence a smaller economy.Second,a major default is likely to have substantial ripple effects throughout the economy.For example,to the extent that financial institutions hold a considerable amount of government debt in their portfolios,they may be forced into bankruptcy,potentially threatening the entire financial and savings system. These effects do not prevent modern democracies from defaulting.Rather,they reveal that there are political costs from default,suggesting that only when the value of default is particularly large is it likely to be considered an option.48 This calculus does not depend on a special political role for bondholders but on the costs of default for nonbondholders.For political institutions to serve as a solu- tion to the sovereign debt problem,effective representation need not be limited to those who hold public debt.What is crucial is that representation cover those who have a stake in the repayment of debt,a group that is generally much larger in contemporary economies.49 Liberal institutions typically impose additional constraints on default.As noted above,one way in which reputational mechanisms failed to control sovereigns in medieval and early modern Europe was that they could pursue"divide and con- quer"strategies,allowing sovereigns to renege successfully on one group while rais- ing money from another.30 To pursue a divide-and-conquer strategy,the sovereign must be able to discriminate finely among categories of lenders.It is precisely here that the universality rules of liberal polities-namely,institutions requiring that like individuals be treated alike-have an important effect.Debt contracts in such sys- 46.Alesina 1988. 47.See Alesina et al.1992. 48.Furthermore,economists have developed an empirical literature studying the likelihood of de- fault in modern developed democracies.Using empirical measures of risk assessed by the market,they are able to investigate the circumstances under which modern democracies risk default.Results from studies of the developed West since World War II suggest that the debt to GNP ratio must be particu- larly large,over 0.5,for default to become a noticeable risk.See Alesina et al.1992.For countries, such as the United States,which have remained well below that level,the value of default has re- mained below the costs. 49.Notice that this argument will be weaker in the case of developing countries whose debt is held almost entirely by foreigners.In such countries,the constituencies in favor of repaying debt can be less potent politically,especially in times of severe financial crisis,when the"ripple effects"of default discussed above are likely to seem small. 50.This principle is modeled in Greif,Milgrom,and Weingast 1994.Veitch 1986 presents evidence of this behavior
Institutional Foundations of Financial Power 13 In modern democracies, a much wider range of interests is represented. There is nothing inherent in such systems to prevent nondebtholders from seeking to default, especially when the debtholders are not an important constituency of the party in power. Nonetheless, default is costly to the rest of the economy and hence to other important constituencies such as labor or nondebtholding capitalist^.^^ Economists list several such costs, and two are relevant for us?' First, default implies that debt finance will be more difficult in the future. Any future increments to spending are therefore likely to require new taxes. Moreover, the demise of debt finance removes the state's ability for tax smoothing, implying lower economic growth and hence a smaller economy. Second, a major default is likely to have substantial ripple effects throughout the economy. For example, to the extent that financial institutions hold a considerable amount of government debt in their portfolios, they may be forced into bankruptcy, potentially threatening the entire financial and savings system. These effects do not prevent modern democracies from defaulting. Rather, they reveal that there are political costs from default, suggesting that only when the value of default is particularly large is it likely to be considered an option.48 This calculus does not depend on a special political role for bondholders but on the costs of default for nonbondholders. For political institutions to serve as a solution to the sovereign debt problem, effective representation need not be limited to those who hold public debt. What is crucial is that representation cover those who have a stake in the repayment of debt, a group that is generally much larger in contemporary ec0nomies.4~ Liberal institutions typically impose additional constraints on default. As noted above, one way in which reputational mechanisms failed to control sovereigns in medieval and early modern Europe was that they could pursue "divide and conquer" strategies, allowing sovereigns to renege successfully on one group while raising money from another.50 To pursue a divide-and-conquer strategy, the sovereign must be able to discriminate finely among categories of lenders. It is precisely here that the universality rules of liberal polities-namely, institutions requiring that like individuals be treated alike-have an important effect. Debt contracts in such sys- 46. Alesina 1988. 47. See Alesina et al. 1992. 48. Furthermore, economists have developed an empirical literature studying the likelihood of default in modern developed democracies. Using empirical measures of risk assessed by the market, they are able to investigate the circumstances under which modern democracies risk default. Results from studies of the developed West since World War I1 suggest that the debt to GNP ratio must be particularly large, over 0.5, for default to become a noticeable risk. See Alesina et al. 1992. For countries, such as the United States, which have remained well below that level, the value of default has remained below the costs. 49. Notice that this argument will be weaker in the case of developing countries whose debt is held almost entirely by foreigners. In such countries, the constituencies in favor of repaying debt can be less potent politically, especially in times of severe financial crisis, when the "ripple effects" of default discussed above are likely to seem small. 50. This principle is modeled in Greif, Milgrom, and Weingast 1994. Veitch 1986 presents evidence of this behavior
14 International Organization tems include fine details such as prioritization and cross-default clauses,limiting the ability of the government to discriminate among bondholders. Institutions of limited government thus serve to constrain political officials and increase the punishments that can be imposed in the event of default.At the same time,these institutions decrease the costs of imposing such penalties.Bulow and Rogoff show that,because a credit boycott hurts lenders as well as the sovereign, the threat of such a boycott lacks credibility:lenders would always be better off renegotiating the loan to get back whatever they can.3 Representative institutions mitigate this problem by providing means of punishing sovereigns,such as elec- toral accountability,that are less costly to implement.All else being equal,then,a state with institutions of the kind identified here will enjoy greater access to credit, and lower rates of interest,than a state whose political leaders are less con- strained.And because access to public debt is a crucial determinant of financial power,such states should enjoy commensurate advantages in international competition. Of course,illiberal states often have access to credit.The theory suggests,how- ever,that these states must pay a premium to do so,and they may face credit rationing.Thus international competition puts a greater strain on their financial resources,both by limiting the amount that can be raised through debt and by increasing their borrowing costs.As a result,illiberal states generally must resort to other,more distortionary mechanisms for raising revenue,such as expropria- tion or printing money.Moreover,this problem gets worse the more unconstrained is the sovereign,as we will see below.In the case of France,limited reforms aimed at increasing the power of debtholders meant that the crown had access to credit, but at higher cost than did its counterpart in Britain.Over the long run,the state could not keep up with its mounting obligations,leading to financial,and ulti- mately political,crises.The Soviet Union represents a more extreme case of a totalitarian state with a command economy that effectively ended its domestic bond market in 1957 and opted instead to rely on direct and indirect expropriation and money creation.Although this strategy permitted the state to raise substantial sums, it proved inferior over the long run,exacerbating the inefficiencies in the econ- omy and helping to undermine efforts at reform.Thus illiberal states can raise money in ways that do not depend on their credibility,but these strategies are distinctly second-best. Case Selection The next two sections apply this logic to two cases of prolonged international competition:the rivalry between Britain and France from 1689 to 1815 and the Cold War between the United States and the Soviet Union.Elsewhere,we demon- strate similar patterns in the seventeenth-century conflict between the Dutch Re- 51.Bulow and Rogoff 1989
14 International Organization tems include fine details such as prioritization and cross-default clauses, limiting the ability of the government to discriminate among bondholders. Institutions of limited government thus serve to constrain political officials and increase the punishments that can be imposed in the event of default. At the same time, these institutions decrease the costs of imposing such penalties. Bulow and Rogoff show that, because a credit boycott hurts lenders as well as the sovereign, the threat of such a boycott lacks credibility: lenders would always be better off renegotiating the loan to get back whatever they can." Representative institutions mitigate this problem by providing means of punishing sovereigns, such as electoral accountability, that are less costly to implement. All else being equal, then, a state with institutions of the kind identified here will enjoy greater access to credit, and lower rates of interest, than a state whose political leaders are less constrained. And because access to public debt is a crucial determinant of financial power, such states should enjoy commensurate advantages in international competition. Of course, illiberal states often have access to credit. The theory suggests, however, that these states must pay a premium to do so, and they may face credit rationing. Thus international competition puts a greater strain on their financial resources, both by limiting the amount that can be raised through debt and by increasing their borrowing costs. As a result, illiberal states generally must resort to other, more distortionary mechanisms for raising revenue, such as expropriation or printing money. Moreover, this problem gets worse the more unconstrained is the sovereign, as we will see below. In the case of France, limited reforms aimed at increasing the power of debtholders meant that the crown had access to credit, but at higher cost than did its counterpart in Britain. Over the long run, the state could not keep up with its mounting obligations, leading to financial, and ultimately political, crises. The Soviet Union represents a more extreme case of a totalitarian state with a command economy that effectively ended its domestic bond market in 1957 and opted instead to rely on direct and indirect expropriation and money creation. Although this strategy permitted the state to raise substantial sums, it proved inferior over the long run, exacerbating the inefficiencies in the economy and helping to undermine efforts at reform. Thus illiberal states can raise money in ways that do not depend on their credibility, but these strategies are distinctly second-best. Case Selection The next two sections apply this logic to two cases of prolonged international competition: the rivalry between Britain and France from 1689 to 1815 and the Cold War between the United States and the Soviet Union. Elsewhere, we demonstrate similar patterns in the seventeenth-century conflict between the Dutch Re- 51. Bulow and Rogoff 1989
Institutional Foundations of Financial Power 15 public and Spain.32 Though the liberal states in these conflicts do not all conform to modern images of democracy,the theory developed here still applies,because it employs a functional definition of democracy,rather than an ideal one based on criteria such as universal suffrage.We raised a problem-the need for credible commitments to repay sovereign debt-and then identified political institutions that help solve this problem.These include representative legislatures with power over budgeting,mechanisms for removing representatives from office,universal- ity norms-institutions,in short,which ensure that debtholders and those with a stake in the repayment of debt have reliable means to punish the sovereign in the event of default.All or most of these institutions tend to be present in democratic polities,and yet,on their own,they are not sufficient to qualify a polity as demo- cratic by current standards.33 In particular,our analysis is silent with respect to the extent of franchise-a factor that separates competitive oligarchies from true democracies.54 Thus,though we may use the term"democracy"as a shorthand, our interest is not in democracy per se,but in a more basic set of institutions.s5 For this reason,it is appropriate to place the twentieth-century United States, eighteenth-century Britain,and seventeenth-century Holland into the same grouping. It must be noted at the outset that there are substantial difficulties to testing this theory.We do not claim,nor do we attempt to prove,that differential access to credit is the sole reason for the pattern of outcomes observed.In a military competition fought over a period of several decades,a large number of factors come into play,including geography,population growth,technological advances, and the quality of military leadership.Moreover,the determinants of economic growth are many and varied.While the efficient use of debt to finance inter- national competition contributes to long-term development,it is by no means wholly responsible for the different growth rates we observe.Given the need to weigh the influence of one factor among many,a small n,case study analysis has inherent limitations. Nevertheless,there is still value in the exercise.Our argument not only posits a causal connection between the domestic political institutions of the rival states and the outcomes of these conflicts but also specifies a number of intervening pro- cesses through which this connection plays out.It suggests that states with repre- sentative government should enjoy superior access to public credit,and that this should translate into fiscal,economic,and military advantages:in particular,the ability to pursue optimal tax-smoothing policies and to sustain high levels of mil- itary expenditures without compromising long-term economic growth.Despite its limitations in dealing with events that may be over-determined,case analysis can corroborate the plausibility of a causal argument by confirming that hypothesized 52.Schultz and Weingast 1998. 53.Doyle 1983;see also Gurr 1990. 54.Dahl1971,7. 55.We are grateful to an anonymous reviewer for this point
Institutional Foundations of Financial Power 15 public and Spain." Though the liberal states in these conflicts do not all conform to modern images of democracy, the theory developed here still applies, because it employs a functional definition of democracy, rather than an ideal one based on criteria such as universal suffrage. We raised a problem-the need for credible commitments to repay sovereign debt-and then identified political institutions that help solve this problem. These include representative legislatures with power over budgeting, mechanisms for removing representatives from office, universality norms-institutions, in short, which ensure that debtholders and those with a stake in the repayment of debt have reliable means to punish the sovereign in the event of default. All or most of these institutions tend to be present in democratic polities, and yet, on their own, they are not sufficient to qualify a polity as democratic by current ~tandards.~~ In particular, our analysis is silent with respect to the extent of franchise-a factor that separates competitive oligarchies from true demo~racies.~'Thus, though we may use the term "democracy" as a shorthand, our interest is not in democracy per se, but in a more basic set of institution^.^^ For this reason, it is appropriate to place the twentieth-century United States, eighteenth-century Britain, and seventeenth-century Holland into the same grouping. It must be noted at the outset that there are substantial difficulties to testing this theory. We do not claim, nor do we attempt to prove, that differential access to credit is the sole reason for the pattern of outcomes observed. In a military competition fought over a period of several decades, a large number of factors come into play, including geography, population growth, technological advances, and the quality of military leadership. Moreover, the determinants of economic growth are many and varied. While the efficient use of debt to finance international competition contribute? to long-term development, it is by no means wholly responsible for the different growth rates we observe. Given the need to weigh the influence of one factor among many, a small iz, case study analysis has inherent limitations. Nevertheless, there is still value in the exercise. Our argument not only posits a causal connection between the domestic political institutions of the rival states and the outcomes of these conflicts but also specifies a number of intervening processes through which this connection plays out. It suggests that states with representative government should enjoy superior access to public credit, and that this should translate into fiscal, economic, and military advantages: in particular, the ability to pursue optimal tax-smoothing policies and to sustain high levels of military expenditures without compromising long-term economic growth. Despite its limitations in dealing with events that may be over-determined, case analysis can corroborate the plausibility of a causal argument by confirming that hypothesized 52. Schultz and Weingast 1998. 53. Doyle 1983; see also Gurr 1990. 54. Dahl 1971, 7. 55. We are grateful to an anonymous reviewer for this point