usa = Greece ?
A challenge. Paul Krugman asks how a country like the USA could have a Greek style crisis.
In his Mundell-Fleming Lecture Krugman defined Greek style crisis twice ” a Greek-style crisis of soaring interest rates” and ” a Greek-style scenario of higher rates and a slump in the real economy” . He says there is no plausible mechanism.
I will ignore the word “plausible”.
First the US Federal government can’t default on dollar denominated debt (if it is considered as a whole including the Treasury, the Federal reserve banks and let’s toss in the Social Security Administration to keep them company). The Fed can create as many dollars as it pleases. The rough equivalent of default is monetization of the debt in which the Fed creates enough dollars to pay it, but the dollar loses its value. Another way to understand this is to note that the US Federal Government can inflate away the present value of payments on long term bonds. Losses to investors can be the same if a bond which pays cents on the dollar because of default and if a bond loses value because payments are discounted at higher market clearing nominal interest rates.
I will use “loss of confidence” to mean loss of confidence that a treasury is solvent so it will be able to pay its debts without default or extraordinarily high inflation. Like Krugman I will consider a loss in such confidence in a treasury which has borrowed in its own currency and will assume that the country in question is in a liquidity trap. I will also assume that fiscal policy can’t be changed quickly, that is that the US Congress is paralyzed (as it is).
I will attempt to avoid Krugman’s conclusion by changing the assumption about the form of the loss of confidence, by considering banks more explicitly and (mainly) by invoking the expected inflation imp.
First, the principal determinant of long-term rates is the expected path of short-term rates. Normally we would only expect a large rise in long rates if investors expect short rates to rise sharply in the future. So we’re back to asking why the central bank would raise short-term rates; a Taylor rule would say that it will do so only if the economy booms or inflation rises (of which more below), and at the zero lower bound it would require a large shock before there would be any change. That is, focusing on the long rate does not seem to change the basic story.
However, it is possible (in theory) for there to be a large jump in expected inflation. In some standard models this can happen in response to a sunspot — a signal which doesn’t affect tastes or technology and so isn’t a shock in the ordinary sense of the word. One of these models include the overlapping generations model with money. But the more relevant class of models are standard New Keynesian models with a Taylor rule with too low a coefficient on lagged inflation.
Formal models don’t rule out such a “large shock”. Experience tells us that they don’t happen in developed countries (at least after the end of the German hyperinflation) but experience also told us that US housing prices don’t fall nationwide and that public and money center bank paid short term interest rates move together (for what it’s worth, I am very confident that such a jump in US expected inflation will not occur — this post is purely an intellectual exercize).
It can be impossible for a monetary authority to achieve low interest rates forever. High expected inflation and a low discount rate lead to hyperinflation and, in the end, to abandonment of the currency (at least this is what happened in Germany). In this case there would be no nominal interest rate on dollar denominated assets, because there would be no dollar denominated assets. I think high expected inflation makes high nominal interest rates inevitable even if the monetary authority is willing to accept high inflation.
An insolvent Treasury doesn’t have to choose between cutting spending, raising taxes or inflating away the debt, there are other options.
One way in which a Treasury in which no one has confidence can cover a deficit is by forcing some entity to buy its debt at a price it dictates. This is, in effect, a tax and would require legislation (recall I assume Congress won’t do anything). Relatedly, the Federal Reserve Board can and does force banks to hold reserves of high powered money — the sum cash in vaults and deposits at federal reserve banks. Each kind of disguised tax is called “financial repression”. To get to a Greek style crisis, I have to rule out financing debt by creating money and avoiding inflation by imposing absurdly high reserve requirements (higher than 100% of deposits I think).
I honestly think such a strategy might be ruled out by the insolvency of banks. In particular, an increase in nominal interest rates can make banks insolvent, since they borrow short term and lend long term. I think the protagonist of this story is the 30 year fixed interest rate mortgage — that is I stress the vulnerability of financial intermediaries which borrow short term and lend long term. In the story (at least) total mortgage debt is roughly equal to US Federal Government debt held by agents other than the Social Security Administration and the Federal Reserve System.
I think the cost of bailing out mortgage lenders could eliminate the windfall from the reduced value of long term public nominal debt.
The point (if any) of this post, is that a sharp increase in expected inflation and nominal interest rates can cause a slump in the real economy, because of the effect on banks. I think this is what Kenneth Rogoff had in mind (it might also be what Larry Summers has in mind in the current discussion).
Further rambling after the jump
Several commentators – for example, Rogoff (2013) — have suggested that a sudden stop of capital inflows provoked by concerns over sovereign debt would inevitably lead to a banking crisis, and that this crisis would dominate any positive effects from currency depreciation. If correct, this would certainly undermine the optimism I have expressed about how such a scenario would play out.
“The question we need to ask here is why, exactly, we should believe that a sudden stop leads to a banking crisis. The argument seems to be that banks would take large losses on their holdings of government bonds. But why, exactly? A country that borrows in its own currency can’t be forced into default, and we’ve just seen that it can’t even be forced to raise interest rates. So there is no reason the domestic-currency value of the country’s bonds should plunge.
First I note that if a loss of confidence in US public debt causes nominal interest rates to rise, banks will take large losses on their long term nominal holdings including mortgages not just on the government bonds. Second, I note that the “can’t even be forced to raise interest rates” is different from “it would require a large shock “. In general existing theory doesn’t usually allow sunspots to cause small shifts in expectations but rule out large jumps in expectations. In particular, a shift to an equilibrium in which foreign currency is used as a means of payment is generally theoretically possible (and any model which rules this out completely must be misleading because such things have happened).
Again I am sure that Krugman is right about the USA, Japan and the UK. But I don’t think there are no existing models in which he is wrong. I think his judgment is based on observing the experiences of large modern developed economies.
There are two more things which I would like to write, but I can’t fit them in so I will just add them here.
Krugman shifted from “loss of confidence” to “sudden stop”. Here I think his background in international economics is relevant. The sudden stop is a sudden reduction in the amount foreign entities are willing to lend to the country under consideration. A loss of confidence in the US Treasury would also cause capital flight as domestic entities send their wealth abroad. The reduction of demand for US public debt could be greater than foreign holdings of US public debt. I think also here, Krugman is allowing experience to affect his theory (as he should). In practice since 1929 or so it has been bankers who suddenly change their minds about risks all at the same time. But in theory, ordinary people could be as prone to manias panics and crashes as bankers.
Second, there is the long run and the abuse of the concept by macroeconomists. The absolutely standard approach (uniting new Classical and new Keynesian macroeconomists) is to assume that there is a unique long run equilibrium and macroeconomic theory concerns convergence to this equilibrium. This is not at all an implication of core assumptions of rationality and intertemporal optimization. It is a standard feature of standard models, because it is generally agreed that models without this property are to be avoided. It is a widely shared assumption not a theoretical result or a stylized fact supported by evidence.
Krugman made the valuable contribution of noting that essentially all of the implications of intertemporal optimization with a unique long run equilibrium can be obtained using the simplest model with two periods — the present and the long run. In fact his ” “miniaturized” New Keynesian model ” has only two periods. He had no need for the moderately impressive mathematics typically used by macroecomists. He focused on the key feature of dynamic macroeconomic models. In this way, he has shown how absolutely central the assumption that in the long run everything will be normal is. In particular, he doesn’t consider hyperinflation, or the fear of hyperinflation, partly because, given his assumptions, hyperinflation would accelerate, so there wouldn’t be a balanced growth path.
I conclude again that I am sure Krugman is right, because I trust his judgment based on his experience (and my experience that he is almost always right). But I am sure of this, because I don’t care whether there is an economic model in which his conclusion would be incorrect.