Money and Debt

For economists

As regular readers will know, my advocacy of helicopter money (HM) does not depend on it being different from, or better (at stimulating demand) than, fiscal policy. [1] So, for example, when Fergus Cumming from the Bank of England said that if after HM the government recapitalised a central bank this “reduces the initial stimulus to a vanilla, bond-financed fiscal transfer”, then that sounds just fine to me. Except, of course, to note that HM is not just like fiscal policy because (a) HM may be quicker to implement than conventional fiscal policy, and speed matters (b) HM can bypasses both genuine debt fears and deficit deceit (c) with HM there is no chance of monetary offset.

Much the same is true for this Vox article by Claudio Borio et al. They argue that if interest is paid on all bank reserves, then HM is “is equivalent to debt-financing from the perspective of the consolidated public sector balance sheet”. Maybe, but why should that be a problem? It is only a problem if you set up a straw man which is that HM has to be more effective than a bond financed helicopter drop.

The reason some people think it is not a straw man is that, if you set up a model where Ricardian Equivalence holds and you have an inflation targeting central bank, a bond financed lump sum tax cut would have no impact. Then you would indeed want HM to do something more. And perhaps it could, if it led agents to change their views about monetary policy. While such academic discussions may be fun, I also agree with Eric Lonergan that “theoretical games being played by some economists, which masquerade as policy insights, are confusing at best.” A good (enough) proportion of agents will spend HM - at least as many as spend a tax cut - for perfectly sound theoretical reasons. [2]

The Bario et al article does raise an interesting question. When the central bank pays interest on all reserves, what is the difference between money and bond financing? Reserves would seem to be equivalent to a form of variable interest debt that can be redeemed for cash at any time. It is exactly the same question raised in a paper by Corsetti and Dedola, an early version of which I discussed here. The answer their model uses is that the central bank would never default on reserves, whereas debt default is always an option.

I think this all kind of misses the point. Base or high powered money (cash or reserves) is not the same as government debt, no matter however many times MMT followers claim the opposite. (For a simple account of why the tax argument is nonsense, see Eric Lonergan here.) Civil servants can frighten the life out of finance ministers by saying that they may no longer be able to finance the deficit or roll over debt because the market might stop buying, but they cannot do the same by saying no one will accept the money their central bank creates. [3] Money is not the government’s or central bank’s liability. (For a clear exposition, see another piece by Eric, or this by Buiter.) Money is not an obligation to make future payments. Money is valuable because, as Eric describes here, it is an established network.

Bario et al seem to want to claim that because central banks nowadays control interest rates by using the interest they pay on reserves, this somehow creates an obligation. Reserves are like variable rate, instant access debt that banks get for nothing.

I think we can see the problem with this line of argument by asking what happens if obligations are broken. If the government breaks its obligation to service or repay its debt we have default, which has extremely serious consequences. If the central bank decides on a different method to control short term interest rates because paying interest on reserves is too much like a transfer to banks, no one but the banks will notice.

So reducing the macroeconomics of helicopter money to fiscal policy is not an argument against it. Furthermore money created by the central bank is not the same as government debt, even if interest is paid on reserves.

[1] They would also know - unlike Jörg Bibow - that I do not think there is any kind of contest between fiscal policy and HM, because the fiscal authority moves first.

[2] The two main reasons some people will spend a tax cut is if they are borrowing constrained, or if they think there is a non-zero probability that the tax cut will be paid for by reducing government spending. An additional reason for spending HM is that it might be permanent if it avoids the central bank undershooting its inflation target.

[3] If the finance minister knows some macroeconomics they would of course realise that not being frightened by the second means you should not be frightened by the first. But that does not negate the conceptual difference.           

Postscript (8/6/16): This by Biagio Bossone provides a very good complement to my analysis, looking a why HM is not 'permanent' and discussing interest on reserves

Fiscal backing

In an earlier post I went through the logic of why we do not think higher government debt necessarily causes inflation, even if that debt is denominated in nominal terms, as long as the central bank does not monetise that debt. As I argued there, talk of monetisation is largely unnecessary: we just need to say that the central bank uses interest rates to control inflation, and can therefore offset the impact of any increase in government debt.

However, as Mervyn King said, central banks are obsessed with budget deficits. This seems to contradict the previous paragraph. Are there some ways in which central banks would either lose the power to control interest rates, or be forced to abandon any inflation targets, as a result of fiscal policy? 

In the previous post the thought experiment I considered was a sustainable increase in the level of government debt. By sustainable I mean that the fiscal authorities raise taxes (or cut spending) to service this higher level of debt. But suppose they do not: suppose the budget deficit increases because spending is higher, but there is no sign that the government is prepared either to cut future spending or raise taxes to a sustainable level.

In 1981 Sargent and Wallace published a well known paper which said that, in this situation, the central bank could in the short term control inflation, but in the longer term inflation would have to rise to create the seignorage to make the government budget constraint balance. In other words, to keep the economy stable the central bank would eventually be forced to monetise. This was later generalised by the Fiscal Theory of the Price Level (FTPL). If the government did not act to stabilise debt itself (which Eric Leeper called – a little oddly - an active fiscal policy, and which others - including Woodford, Cochrane and Sims - have called even more confusingly a non-Ricardian policy [1]), then the price level would adjust to reduce the real value of government debt. Fiscal policy determines inflation.

One of the critiques of this theory is that the government budget constraint appears not to hold at disequilibrium prices. See, for example, Buiter here, and a response from Cochrane. I do not want to go into that now. Let’s also concede that if the monetary authority does either follow a rule that allows the price level to rise (by fixing the nominal interest rate for example), or tries to move interest rates to both stabilise debt and inflation (as in my recent paper with Tatiana Kirsanova), then the FTPL is correct.

The case I want to focus on here is where the central bank refuses to do either of those things, but carries on controlling inflation and ignoring debt. Suppose the government is running a deficit which is only sustainable if we have a burst of inflation which devalues the existing stock of government debt, but the central bank refuses to allow inflation to rise. You can say it does this by fixing the stock of money, or by raising the rate of interest - I do not think it matters which. This is an unstable situation: interest payments on the stock of debt at the low price level can only be paid for by issuing more debt, so debt explodes. In this situation, we have a game of chicken between the government and central bank.

Now the game of chicken would probably end when the markets refused to buy the government’s debt. That would be the crunch moment: either the central bank would bail the government out by printing money, or the government would default, which forces it to change fiscal policy. But in Buiter there is an elegant equilibrium outcome: the market just discounts the value of debt by an amount that allows the central bank to set the price level, but for the government’s budget constraint to hold at that price level. We get partial default. This discount factor becomes the extra variable that solves for the tension that both fiscal and monetary policy are trying to determine the price level.

You could quite reasonably suggest that such a central bank could not exist, because the government has ultimate power. It can always instruct the central bank to monetise the debt. However suppose the central bank actually managed the currency for a whole group of nations, and could only be instructed to do anything if they all agreed to do so. Furthermore that central bank was located in the one country in that group that would never contemplate monetisation, so it would be immune to pressure ‘from the street’. That central bank should be pretty confident it could win any game of chicken. [1]

Has any of this any relevance to today’s advanced economies? It seems to me pretty clear that these governments are not playing any game of chicken. Quite the opposite in fact: they are being far too enthusiastic in doing what they can to stabilise debt, despite there being a recession. So we certainly do not seem to be in a FTPL type world. Instead monetary policy right now retains fiscal backing.

Yet in a way we are having the wrong conversation here. Rather than trying to convince central banks that their fears are groundless, we should be asking whether monetary policy should – of its own free will – raise inflation to help reduce high levels of debt. I agree with Ken Rogoff that it should, and have argued the case here. Yet however optimal such a policy might be, the chances of it happening in today’s environment are nil. It looks like we may have to go through a lost decade before we are allowed to contemplate such things. 

[1] I guess a rationale for calling this fiscal policy ‘active’ is that stable regimes in Leeper require one partner to be active and the other passive. So in the normal regime monetary policy is active and fiscal passive, and this flips in a FTPL regime. In a FTPL regime, Ricardian Equivalence no longer holds (because taxes are not raised following a tax cut) – hence the label non-Ricardian.

[2] In this situation, would buying that government’s debt ‘show weakness’ in the game? If we follow Corsetti and Dedola and treat reserves as default free debt issued by the central bank rather than money, then not at all. Instead the central bank is giving the fiscal authority the best chance it can to put its house in order, by removing any bad equilibrium, but it retains the power to force default at any point. We no longer have Buiter’s method of resolving that game, but only because the central bank has the means which could force a win. As long as the government believes that the central bank would prefer the government to default rather than see inflation rise, the government should back down.

Did Ricardian Equivalence kill the Pigou effect?

For macroeconomists

After the last time the world got into a liquidity trap, there was a debate about whether price flexibility would be sufficient to get us out of the trap. That debate tended to assume a fixed money supply. With that assumption, the answer today would be yes, if falling prices raised inflation expectations (given long run neutrality) and therefore reduced real interest rates. Back then that story was not so popular, perhaps because the debate pre-dated rational expectations. Instead the argument at the time focused on the Pigou or Real Balance effect. Falling prices raised the value of outside money, so everyone would feel wealthier and spend more.

We do not hear this argument so much nowadays. I have not seen this discussed in the advanced textbooks I know well (for example neither term is in the index of Romer or Obstfeld and Rogoff), so I was wondering why that was. Is the Pigou effect not what it was once thought to be? I could not find a clear answer to this question anywhere, but of course that may be my failing. So here are my thoughts, but they come with the possibility that I have just missed something. If I have, I will rewrite the post accordingly.

What I did find were a few papers that appeared to suggest that Ricardian Equivalence (REq) killed the Pigou effect. Here is a quote from a paper by Peter Ireland. After talking about REq, he writes

“Less widely appreciated, however, is a closely related finding, presented most explicitly by Weil (1991) but also implicit in earlier work by Sachs (1983) and Cohen (1985). These authors show that government-issued fiat money will not be perceived as a source of private-sector wealth if the households owning that money are the same households that, first, receive all of the transfers or pay all of the taxes associated with future changes in the money supply and that, second, incur all of the opportunity costs associated with carrying the money stock between all future periods. We are used to the idea of Ricardian Equivalence implying that government debt is not net wealth. Essentially consumers internalise the government’s budget constraint. But that argument applies to outside money as much as government debt. We can replace initial values of debt and money by the discounted future stream of primary surpluses they support.”

The easiest way to describe REq is that the infinitely lived representative consumer consolidates the government’s intertemporal budget constraint (IBC) into its own. Suppose this consumer owns some nominal (non-indexed) government debt, and the price level falls. Is that consumer better off? The real value of the future interest they receive on that debt will be higher, but this will be offset by the higher taxes in real terms that the government will raise to pay for this. The same argument applies to the higher real redemption value of the debt.

Ireland argues that exactly the same points can be made about outside money. Suppose money pays no interest, but consumers hold it because of the liquidity services it provides.

But if the consumer already has all the liquidity services they need (as they do in a liquidity trap), a fall in prices that creates more of this asset in real terms does not make the consumer better off on this account. So what about the redemption value of the additional real balances?

Here I’m inclined to think that money is different from government debt. In a paper[1] that I do not think has been published, Willem Buiter argues that money is irredeemable. The government only promises to redeem money with itself. So if I get a tax cut that is financed by printing money rather than issuing debt, there is no offsetting future tax liability. For this reason, money – unlike government debt – is net wealth for the consolidated public and private sectors.

Now a standard response is to say that a money financed tax cut does not make the consumer better off because the price level will rise, reducing the purchasing power of that money. It seems to me that is a different argument to REq – it requires going beyond just thinking about budget constraints. It is also an argument that does not apply to the Pigou effect, which is what happens if prices fall, raising the value of real balances.

Does the irredeemable nature of money rescue the Pigou effect from the REq argument? Yes and no. There is a crucial difference between Buiter’s analysis and the traditional view. In Buiter, it is the present discounted value of the terminal stock of base money that is net wealth for the consolidated private and public sectors, rather than its current value. To see why this matters, consider the liquidity trap case again.

As we have already noted, there is no liquidity trap in the flexible price case when the government holds the nominal stock of money constant, because falling prices today imply higher expected inflation. We do not need a Pigou effect. But the more interesting case, which I have talked about before, is where the government or central bank has an inflation target. In this case the authorities prevent inflation expectations rising, so real interest rates do not fall.

In that case nominal money will not be held constant when prices fall. Instead, the authorities will contract the nominal money stock in line with falling prices, to make sure inflation does not rise. As a result, there will be no increase in consumption, because the terminal value of nominal money falls, and its real value stays constant. Or, to put the same point another way, higher future taxes required to reduce the money stock will offset the wealth impact of higher current real money balances. There is no Pigou effect.

This is all terribly stylised and unrealistic, so there is no need to add comments that just point this out. However, I hope I’m not the only one who thinks this thought experiment is 
interesting. I also think that the proposition that inflation targets prevent macroeconomic ‘self-correction’ even when prices are flexible has a symbolic importance.

[1] Buiter, W.H. (2003) Helicopter Money: Irredeemable Fiat Money and the Liquidity Trap, NBER Working Paper No. 10163.