Institutional Foundations of Financial Power 11 ability of potential lenders to sanction him for default,the sovereign benefits from an increase in the penalties that can be imposed on him. It might seem that reputational considerations alone would be sufficient to in- duce the sovereign to honor his commitments.After all,once the sovereign de- faulted,lenders would think twice before extending further loans.The possibility of a credit boycott would then create a strong incentive for the sovereign to repay his debts.As a number of writers have suggested,however,this reputational mech- anism is insufficient to ensure that the sovereign will honor his agreements.37 Two major obstacles hinder the lending community's ability to police the sov- ereign.First,credit boycotts are difficult to organize and sustain.Because the sov- ereign is unlikely to renege on all of his creditors at once,their interests will be divided,and the sovereign will be able to play some off against others.At the same time,creditors face the usual free-rider problems associated with this kind of collective action,as there would be significant incentives to defect from a boy- cott in order to become the state's sole source of credit.38 Second,a credit boycott is costly to the lenders themselves,because they must forgo their source of liveli- hood.As Bulow and Rogoff demonstrate,this puts lenders in a bad bargaining position,allowing the sovereign to force them to accept less attractive terms than those originally agreed to.3 In addition,reputational mechanisms only work when the actor places suffi- cient weight on future costs.If the sovereign discounts the future quite heavily, the threat of a credit boycott may have little impact on his calculus.A sovereign is likely to have a short time horizon in times of war or major crises,precisely when his need for credit will be greatest.40 This does not mean that reputational consid- erations will never cause the sovereign to honor his debts-only that reputation alone is unreliable. In practice,this difficulty rations the amount of credit available to the sovereign and often raises the costs of borrowing.4 Lending money to someone who cannot be forced to repay is a risky business,so creditors demand a"risk premium"in the form of higher interest rates.Ironically,then,the sovereign's unfettered power makes it quite costly for him to raise money through voluntary loans. Limited Government as a Commitment Technology The institutions of limited government provide a solution to this problem.Institu- tions constrain individuals'actions by shaping the incentives they face.42 In par- ticular,institutional arrangements can modify the incentives of the sovereign by 37.Alesina et al.1992;see also Bulow and Rogoff 1989;Greif,Milgrom,and Weingast 1994;and Veitch 1986. 38.Weingast 1997a. 39.Bulow and Rogoff 1989. 40.North 1981,chap.11. 41.We provide a model incorporating many of these details in a companion article,Schultz and Weingast 1998. 42.North1990.Institutional Foundations of Financial Power 11 ability of potential lenders to sanction him for default, the sovereign benefits from an increase in the penalties that can be imposed on him. It might seem that reputational considerations alone would be sufficient to in￾duce the sovereign to honor his commitments. After all, once the sovereign de￾faulted, lenders would think twice before extending further loans. The possibility of a credit boycott would then create a strong incentive for the sovereign to repay his debts. As a number of writers have suggested, however, this reputational mech￾anism is insufficient to ensure that the sovereign will honor his agreement^.^' Two major obstacles hinder the lending community's ability to police the sov￾ereign. First, credit boycotts are difficult to organize and sustain. Because the sov￾ereign is unlikely to renege on all of his creditors at once, their interests will be divided, and the sovereign will be able to play some off against others. At the same time, creditors face the usual free-rider problems associated with this kind of collective action, as there would be significant incentives to defect from a boy￾cott in order to become the state's sole source of credit.38 Second, a credit boycott is costly to the lenders themselves, because they must forgo their source of liveli￾hood. As Bulow and Rogoff demonstrate, this puts lenders in a bad bargaining position, allowing the sovereign to force them to accept less attractive terms than those originally agreed to.39 In addition, reputational mechanisms only work when the actor places suffi￾cient weight on future costs. If the sovereign discounts the future quite heavily, the threat of a credit boycott may have little impact on his calculus. A sovereign is likely to have a short time horizon in times of war or major crises, precisely when his need for credit will be greatest?' This does not mean that reputational consid￾erations will never cause the sovereign to honor his debts-only that reputation alone is unreliable. In practice, this difficulty rations the amount of credit available to the sovereign and often raises the costs of borrowing?' Lending money to someone who cannot be forced to repay is a risky business, so creditors demand a "risk premium" in the form of higher interest rates. Ironically, then, the sovereign's unfettered power makes it quite costly for him to raise money through voluntary loans. Limited Government as a Commitment Technology The institutions of limited government provide a solution to this problem. Institu￾tions constrain individuals' actions by shaping the incentives they face?2 In par￾ticular, institutional arrangements can modify the incentives of the sovereign by 37. Alesina et al. 1992; see also Bulow and Rogoff 1989; Greif, Milgrom, and Weingast 1994; and Veitch 1986. 38. Weingast 1997a. 39. Bulow and Rogoff 1989. 40. North 1981, chap. 11. 41. We provide a model incorporating many of these details in a companion article, Schultz and Weingast 1998. 42. North 1990