The NYT Room for Debate on the S&P downgrade includes some posts that I think contain some really misleading statements. The term "room for debate" is rather inapt, since we write on our own, with no opportunity to comment on others' posts.
The argument that a country with a fiat currency cannot default on its debt is perhaps the most egregious example. As Megan McArdle points out, the U.S. can at best inflate away its past obligations. But its future obligations are the ones that are crushing.
The government cannot keep unlimited promises to future recipients of entitlements. If you think that fiat money changes this, then consider what happens when national output is $100 and promises to seniors are $200. In that case, no matter how much money you print, you cannot give seniors more than $100 in output.
If you define government obligations as debt that has already been issued by the Treasury, then the government could cover those obligations by printing money. However, once you include future obligations, the ability to print money does not in any way guarantee the ability to meet obligations.
Catherine Rampell seems to agree with my point that debt crises are not predictable by regulatory ratings or other indicators.
Jim Manzi agrees with one of my other points, which is that the politics of fiscal policy may change in a crisis. He writes,
If you think about it, any real solution to the federal deficit problem is currently politically impossible; yet we know mathematically that, barring a productivity miracle, the situation cannot persist indefinitely. Therefore, we know that some change that currently seems politically impossible is all-but-certain to happen sooner or later.