CAMBRIDGE – A year after the United States government allowed the investment bank Lehman Brothers to fail but then bailed out AIG, and after governments around the world bailed out many other banks, key question remains: when and how should authorities rescue financial institutions?
It is now widely expected that, when a financial institution is deemed “too big to fail,” governments will intervene if it gets into trouble. But how far should such interventions go? In contrast to the recent rash of bailouts, future government bailouts should protect only some creditors of a bailed-out institution. In particular, the government’s safety net should never be extended to include the bondholders of such institutions.
In the past, government bailouts have typically protected all contributors of capital of a rescued bank other than shareholders. Shareholders were often required to suffer losses or were even wiped out, but bondholders were generally saved by the government’s infusion of cash.
For example, bondholders were fully covered in the bailouts of AIG, Bank of America, Citigroup, and Fannie Mae, while these firms’ shareholders had to bear large losses. The same was true in government bailouts in the United Kingdom, Continental Europe, and elsewhere. Bondholders were saved because governments generally chose to infuse cash in exchange for common or preferred shares – which are subordinate to bondholders’ claims – or to improve balance sheets by buying or guaranteeing the value of assets.
A government may wish to bail out a financial institution and provide protection to its creditors for two reasons. First, with respect to depositors or other creditors that are free to withdraw their capital on short notice, a protective government umbrella might be necessary to prevent inefficient “runs” on the institution’s assets that could trigger similar runs at other institutions.
Second, most small creditors are “non-adjusting,” in the sense that they are unable to monitor and study the financial institution’s situation when agreeing to do business with it. To enable small creditors to use the financial system, it might be efficient for the government to guarantee (explicitly or implicitly) their claims.
But, while these considerations provide a basis for providing full protection to depositors and other depositor-like creditors when a financial institution is bailed out, they do not justify extending such protection to bondholders.
Unlike depositors, bondholders generally are not free to withdraw their capital on short notice. They are paid at a contractually specified time, which may be years away. Thus, if a financial firm appears to have difficulties, its bondholders cannot stage a run on its assets and how these bondholders fare cannot be expected to trigger runs by bondholders in other companies.
Moreover, when providing their capital to a financial firm, bondholders can generally be expected to obtain contractual terms that reflect the risks they face. Indeed, the need to compensate bondholders for risks could provide market discipline: when financial firms operate in ways that can be expected to produce increased risks down the road, they should expect to “pay” with, say, higher interest rates or tighter conditions.
But this source of market discipline would cease to work if the government’s protective umbrella were perceived to extend to bondholders. If bondholders knew that the government would protect them, they would not insist on getting stricter contractual terms when they face greater risks. The problem of “moral hazard” – which posits that actors will take excessive risks if they do not expect to bear fully the consequences of their actions – is commonly cited as a reason not to protect shareholders of bailed-out firms. But it also counsels against protecting firms’ bondholders.
Thus, when a large financial firm runs into problems that require a government bailout, the government should be prepared to provide a safety net to depositors and depositor-like creditors, but not to bondholders. In particular, if the firm’s equity capital erodes, the government should not provide funds (directly or indirectly) to increase the cushion available to bondholders. Rather, bonds should be at least partly converted into equity capital, and any infusion of new capital by the government should be in exchange for securities that are senior to those of existing bondholders.
Governments should not only avoid protecting bondholders after the fact, when the details of a bailout are worked out; but should also make their commitment to this approach clear in advance. Some of the benefits of a government policy that induces bondholders to insist on stricter terms when financial firms take larger risks would not be fully realized if bondholders believed that the government might protect their interests in the event of a bailout.
In other words, governments should establish bailout policies before the need to intervene arises, rather than make ad hoc decisions when financial firms get into trouble. The best policy should categorically exclude bondholders from the set of potential beneficiaries of government bailouts. This would not only eliminate some of the unnecessary costs of government bailouts, but would also reduce their incidence.