John McHale over at Irish Economy has written a post on the possibility of a default by a sovereign currency issuer. In the post he discusses Paul Krugman’s stripped-down Mundell-Fleming model in which Krugman shows that a country issuing a sovereign currency not only cannot default but if there is a run on the government bonds of this country the result will be devaluation which will then lead to an expansion of output as exports increase.
Krugman’s model — put together in late 2012 — emulates the argument that MMT and other Post-Keynesian economists have been making for years: if a country has a sovereign currency then the government cannot default and any loss of confidence in the government bonds of that country will only affect the exchange rate. Before dealing with the model itself I want to first deal with some points made by McHale which strike me as misguided.
The first is that McHale seems to think that Krugman’s result only holds in a so-called liquidity trap. But this is simply not true. Even in “normal” times — whatever those might be — the central bank sets a target rate of interest (represented in Krugman’s model by a Taylor Rule). If there is a run on bonds the central banks will not change the rate of interest unless inflation rises. So, even in “normal” times the effects will largely be the same: the central bank will soak up bonds offloaded by foreign investors and all the effects will fall on the currency which, ceteris paribus, will depreciate.
The second point McHale makes that is problematic is the following,
So what then is left out? I think one problem comes with the assumption that that the central bank can still set the domestic interest rate following its usual monetary policy rule (constrained by the zero lower bound). What the stripped down model leaves out is that central bank is only directly affecting the overnight lending rate between banks, and also possibly future expectations of this rate.
This is where the experience of Ireland and other peripheral euro zone economies is relevant. Stressed banks cannot access funding at the central bank’s policy rate. This is what has led to the fragmentation of bank funding markets across the euro zone. The weak creditworthiness of the sovereign is one of the factors affecting the creditworthiness of the banks, as the sovereign remains the primary capital backstop to the banking system. And the high funding costs of the banks is keeping up lending rates to the real economy.
This passage seems to me rather confused for two reasons. First of all, the assertion that “weak creditworthiness of the sovereign is one of the factors affecting the creditworthiness of the banks”. This is true to some extent. But in order for this to affect the creditworthiness of the banks the value of bonds has to fall. In Krugman’s model, however, when bonds are offloaded by foreign investors they are soaked up by the central bank. Thus their nominal value does not fall. Given that the repo operations the central bank engages in rely on this nominal value, the fall in the real value of these bonds (through the fall in the exchange rate) will not make any difference to the bank’s abilities to get funding.
Tied to this McHale again seems to be confusing a sovereign currency issuer with a user of a currency. In the case of the former if there is a funding problem with banks the central bank will step in as lender of last resort. The European situation that McHale refers to only arose because in the Eurozone the ECB was reticent about acting as a lender of last resort. In the US and the UK after the crisis the central banks stepped in to allow stressed banks access to funding. So even in the case where an offloading of bonds by foreign investors did affect the creditworthiness of banks — which it would not — the central bank would use its money creation powers to fill the gap.
So, does that mean that Krugman’s model is correct? No, it is wrong. Even on its own terms Krugman has missed something in his model, as has McHale. The key to this is in the Taylor Rule that Krugman uses to determine the interest rate.
As is well-known the Taylor Rule sets interest rates in line (a) with inflation and (b) with unemployment (some variations leave out the latter, but we can ignore this for our present argument). Now, a devaluation of the currency has two effects: it makes exports cheaper and imports more expensive. The latter effect can, of course, cause price inflation. So even in Krugman’s own model there is a possibility that a fall in the exchange rate would lead to higher inflation which would, in turn, lead the central bank to hike interest rates to meet their Taylor Rule.
Am I saying that I agree with this interpretation? No, because I don’t think that central banks set the interest rate in line with a Taylor Rule at all. I also don’t think that exchange rate movements have linear effects on imports and exports — which is why I would be sceptical about the increased output that Krugman assumes from the devaluation. Regardless, however, Krugman has missed a key component of his own model because he ignored the effects that price rises due to rising import costs might have on the interest rate through his Taylor Rule function.
The lesson? Better to throw this whole New Keynesian approach out altogether. It merely leads to confusion. The Post-Keynesian and MMT approach is far more informative. And what’s more they arrived at Krugman’s results years before him.