Tyler Cowan thinks that this post by Paul Krugman on long term interest rates and a follow up by Brad DeLong are critically important and "two of the best recent economics blog posts, in some time".
Krugman's post deals with the question of why some economists in the administration are concerned that further increases in deficit financing could cause long term rates to spike:
Well, what I hear is that officials don’t trust the demand for long-term government debt, because they see it as driven by a “carry trade”: financial players borrowing cheap money short-term, and using it to buy long-term bonds. They fear that the whole thing could evaporate if long-term rates start to rise, imposing capital losses on the people doing the carry trade; this could, they believe, drive rates way up, even though this possibility doesn’t seem to be priced in by the market.
What’s wrong with this picture?
First of all, what would things look like if the debt situation were perfectly OK? The answer, it seems to me, is that it would look just like what we’re seeing.
Bear in mind that the whole problem right now is that the private sector is hurting, it’s spooked, and it’s looking for safety. So it’s piling into “cash”, which really means short-term debt. (Treasury bill rates briefly went negative yesterday). Meanwhile, the public sector is sustaining demand with deficit spending, financed by long-term debt. So someone has to be bridging the gap between the short-term assets the public wants to hold and the long-term debt the government wants to issue; call it a carry trade if you like, but it’s a normal and necessary thing.
Now, you could and should be worried if this thing looked like a great bubble — if long-term rates looked unreasonably low given the fundamentals. But do they? Long rates fluctuated between 4.5 and 5 percent in the mid-2000s, when the economy was driven by an unsustainable housing boom. Now we face the prospect of a prolonged period of near-zero short-term rates — I don’t see any reason for the Fed funds rate to rise for at least a year, and probably two — which should mean substantially lower long rates even if you expect yields eventually to rise back to 2005 levels. And if we’re facing a Japanese-type lost decade, which seems all too possible, long rates are in fact still unreasonably high.
Still, what about the possibility of a squeeze, in which rising rates for whatever reason produce a vicious circle of collapsing balance sheets among the carry traders, higher rates, and so on? Well, we’ve seen enough of that sort of thing not to dismiss the possibility. But if it does happen, it’s a financial system problem — not a deficit problem. It would basically be saying not that the government is borrowing too much, but that the people conveying funds from savers, who want short-term assets, to the government, which borrows long, are undercapitalized.
And the remedy should be financial, not fiscal. Have the Fed buy more long-term debt; or let the government issue more short-term debt. Whatever you do, don’t undermine recovery by calling off jobs creation.
What Krugman seems to be advocating is the following: if long term rates should start to rise, the Treasury should finance the deficit by issuing more short-term (and less long-term) debt, thereby flattening the yield curve and holding long term rates low. This would prevent capital losses for carry traders (although it would lower the continuing profitability of the carry trade if short rates rise).
In effect, Krugman is arguing that the Treasury should itself act like a carry trader: rolling over short term debt to finance a long-term structural deficit. But why is this not being done already? Take a look at the current Treasury yield curve:
What is currently preventing the Treasury from borrowing at much more attractive short rates to finance the deficit? Is it is a fear of driving up short rates? And if so, won't the same concerns be in place if long term rates start to rise?