In a comment to a post of Brad Setser's, DF challenges (perhaps rhetorically), "What is 'too high'? [Debt-to-GDP is] now 'higher' than in 1929... Let us know, if you have a solution to keep the debt level in a permanent high plateau, or have the debt rise forever, or reduce debt level without any harm for the economy."
I'm glad to oblige:
Sequester the debt. Let it be purchased by entities that are willing to roll it over forever, that will absolutely never sell or demand cash for the debt that they hold, and that will accept further debt in lieu of interest payments. Private actors seem unlikely candidates for permanent debt sequestration, they do want eventual cash flow from the assets they hold.
The central bank of a country that issues debt is the ideal agent for debt sequestration: It never requires cash flow, and its primary goal is to prevent disruption of the economy it helps to regulate. This form of sequestration is called "monetization of debt", and is indefinitely stable. The only downside is that purchases of debt by a central bank may lead to inflation. To the degree central bank debt purchases can be managed without inflation (perhaps due to disinflationary forces elsewhere in the economy), the debt-issuer's central bank is the ideal sequesterer of debt.
Foreign central banks can also serve to sequester debt, but they are "leaky". There are two risks associated with foreign central bank sequestration: 1) Sequestration by a foreign central bank may not be permanent — the debt-issuing economy bears the risk of a disruptive change of policy by a foreign central bank; 2) Foreign central banks may expect cash-flow in the form of interest payments on foreign debt, even if they are willing to permanently roll-over principal. (Cash-flow to domestic central banks is usually an accounting fiction, since a central bank's profits are remitted to the debt-issuer.)
Any level of debt-to-GDP is sustainable with sufficient sequestration. Any amount or rate of debt expansion is sustainable so long as there are actors willing to continue to sequester more debt. However, the greater the stock of sequestered debt, and the greater the rate of debt expansion, the greater the risks of a change in policy, especially by a foreign central bank.
These are the risks Roubini and Setser have long highlighted with respect to the United States' deficits. However, the willingness and capacity of Asian and petrostate central banks to sequester US debt have flown in the face of most people's intuitive limits of plausibility.
At this point, it's not enough to say "wow, that's a lot of debt". Debt-to-GDP can approach infinity, notwithstanding whatever the level was in 1929. If you want to say that things will change due to supply-side constraints, you've got to explain why the capacity of debt-sequesterers will be exhausted, or why their policies are likely to change.
(crossposted from here.)
