jakke:

Wow, yeah, creating more US$ and using them to pay creditors would solve the debt ceiling problem pretty neatly. (Except for the likely steep devaluation of the dollar and the fall in demand for US Treasury bonds and all the fun knock-on effects there, obviously.)

But wouldn’t the Fed have to agree to do this? Like, it’s not like the Treasury can order the Fed to just print indefinitely many dollars. And so far it sounds like they are totally opposed to doing that - or at least Bernanke is, but I can’t imagine any of them thinking it would be an optimal outcome.

Although really if the Fed were planning to do this they’d have to keep it totally secret, or else markets would panic and Congress would chillax indefinitely about the whole debt ceiling thing. So maybe they’re planning to fire up the money photocopiers on August 3 just in case?

Just to clarify, I wasn’t trying to imply that it was an optimal solution to the debt problem, only that it was an alternative to default. In addition, I do believe it is a better alternative than default in the short term. A default has lasting consequences. A short term monetization has significantly less impact. A long term monetization pretty much guarantees a hyperinflationary collapse, which is worse than a default in almost every scenario. Pick your poison, I guess.

You are correct, the Treasury doesn’t really have the power to force the Fed to act this way. Congress does, as the Federal Reserve act specifically delegates Congressional authority over the money system to the Fed. It just requires an act of Congress. But that avenue would require Congress to agree on something, which is unlikely since it’s a tacit admission that they won’t increase the debt ceiling.

I don’t see anyone as needing to force the Fed to act this way. The Fed’s primary purpose is to ensure stability of the banking system. I think it’s pretty obvious that exchanging zero-term liabilities for short-term liabilities is significantly better for banking stability than allowing the treasury market to implode. Once Congress sets the banking system on fire, the Fed is going to find every hose available to put the fire out before letting it burn to the ground.

I see 3 mechanisms the Fed can use (all of which effectively reduce to the same thing, but fall under different Fed powers):

The first is to simply delete US debt held by the Fed. This is due to a rather complicated distinction of what actually is an asset and what is actually a liability for the Fed vs Treasury. Treasuries held on the Fed’s balance sheet are liabilities of the Treasury and assets of the Fed. Dollars are actually Fed liabilities (not Treasury liabilities behaving like Fed liabilities). Note that coins are still liabilities of the Treasury, so even this is an awkward distinction. Where the awkwardness really comes into play is the fact that money production (liabilities of the Fed) does not work in lockstep with the assets of the Fed. For things like QE2 or POMO, it does (the money is created at the time of the Fed purchase of the treasuries, and the Fed issues a liability in exchange for an asset, much like you might get a loan to buy a car). For treasury auctions, it is close (the money is created just before the auctions, which leads to odd personal finance analogies due to fractional reserve banking — the money is used by both the US government to pay its obligations and by the primary dealer to buy the debt). For interest the Fed pays on reserves it is unrelated (the money comes from nowhere, which is analogous to you taking out a loan to buy nothing and then finding out that the money you borrowed never existed and you still need to pay back the loan). For most monetary theory, you can treat dollars as treasuries of zero maturity. And just for more awkwardness, any annual profit made by the Fed is paid to the Treasury (in dollars).

The second mechanism is to destroy equal amounts of dollars and treasuries on the Fed’s balance sheet. The Fed is free to magically create dollars (which seems stupid, since they are a liability of the Fed—it’s like getting a loan, not taking the money, but agreeing to pay back the loan). It can then pay itself the face value of the treasuries and have both go poof. As an intermediate step, the Fed may have to issue treasuries expiring immediately for those that have longer maturities.

The third mechanism is to simply monetize the debt. That effectively means printing money to buy treasuries as they mature. However, the Fed can offset this new money by selling longer duration treasuries that it currently holds on its balance sheet. The new money doesn’t circulate. Again, as an intermediate step, the Fed can buy longer duration treasuries and issue shorter duration treasuries to artificially move the maturity date closer. This is standard POMO stuff. The fact that the Fed gets to willy-nilly issue and buy back treasuries via POMO is why it’s sometimes convenient to think of both treasuries and dollars as liabilities of the singular Fed/government unit with different maturities and interest rates (cash having zero maturity and zero interest … except for excess reserves, which have zero maturity and currently have non-zero interest).

If the Fed didn’t already have a lot of treasuries on its balance sheet, these 3 different mechanisms wouldn’t be effectively equivalent (two wouldn’t be possible). But since the Fed has more treasuries than China, this isn’t much of an issue.

Also, in each of my three scenarios, no new dollars need to circulate in the economy and no net new treasury debt needs to circulate among the market. The net effect is simply a reduction of the treasuries on the Fed’s balance sheet.

(Source: dyn.politico.com)