This prompted by twitter commentary on Buiter’s recent paper explaining why helicopter drops will always work and stimulate the economy; a formal paper echoing the call for money printing by Mark Blyth and Eric Lonergan in Foreign Affairs.
Buiter’s paper conducts surgery on the modern monetary DSGE model. The surgery appears to make the model more realistic. In the original, it’s assumed that somewhere off in the infinite recesses of time, the government will pay back all outstanding bonds it has issued, but will also redeem money it printed for real consumption goods too. In this way it’s policies are set to satisfy the public sector intertemporal budget constraint. Buiter says: no one seriously believes the money is going to be ‘redeemed’. So let’s remove that assumption. His model retains the rest of the original machinery, notably, the ASSUMPTION that for some reason unspecified, holders attach a non-pecuniary benefit to holding money.
In the old model, if the helicopters rained down government bonds on the populus, said populus would think: ‘hang on, these are not net wealth to me [HT Barro], because sooner or later they are going to be paid back out of my taxes. Similarly, in the old model, if the helicopters rained down money, people would say ‘the government is going to have to pay for this gift when it redeems the money from whoever has it, and I will be taxed for it’. In the new model, Buiter makes money net wealth in a way that it wasn’t previously.
Actually, that wasn’t the whole story. In the old model, what people would say when money rained down on them would depend on whether we were at the zero bound or not. If we were, at that point it would be the case that people already had sufficient money that the non-pecuniary like they had for it would have been exhausted. This benefit is assumed to capture money’s function as a convenient medium of exchange. And the idea is that beyond some point, it’s useless to keep your wealth in money [perhaps even costly]. Once that benefit disappears, interest rates have to be zero: this measures the opportunity cost of holding money, which has to be equal to its benefit. If you are still with me, this means that whether money has any effect or not will hinge on whether at the zero bound people view it as net wealth. And if they don’t, because it is to be redeemed, they won’t spend as though it were. This is how Buiter gets his stimulatory effect.
This sounds like a triumph. A more realistic model, showing how helicopter drops can stimulate? Not so fast. Forget the real world for a moment. In the model world, we have just assumed that people value money for its own sake [to capture a parable about how they would value its convenience]. And as modellers, this seems ok either because we feel, or from other work we can derive, that this convenience value will emerge attached to an asset that is redeemed. Once we take away the redemption assumption, it begs the question why the convenience value as a medium of exchange is there. After all, the real story of paper money was that it was once redeemable, and that was why it became an acceptable medium of exchange. So, as modellers, we ought to feel very uncomfortable about Buiter playing fast and loose with the government’s intertemporal budget constraint. [Actually, some modellers, the ‘new monetarists’ like Williamson, Wright, Lagos, Kiyotaki, go ballistic about this mainstream model of money for the simple assumption of convenience value regardless of whether money is in the government budget constraint or not].
You might say: but what about the real world? Surely no-one thinks money is going to be redeemed do they? Surely despite this they will always place a convenience value on holding money? Well, maybe, or maybe not. Maybe once helicopters rain down money, people will think that the UK or US governments will get a taste for it, will think that we have become ‘like Zimbabwe’ and run for gold, or other metals, or even, as in revolutionary France as their currency got debased, wine and cheese. What about the law of legal tender? Someone once said to me in the Bank of England that the reason why money has value is because of the law of legal tender. [Which stipulates that a debt is to be considered closed if settled with, legal tender]. This has a grain of truth in it, but a lot of falsehood. The law of legal tender, and the widespread use and acceptance of pound coins [first draft wrote pound notes, revealing my age] are symbiotic facets of our monetary system, and the former doesn’t solely determine the latter. I’ve no doubt that if the government started debasing the currency, dollarisation would ensue here as fast as could be, and the law of legal tender would not matter at all.
This leads me to be strongly against helicopter drops of money. I don’t buy the model-based explanation, because Buiter’s removal of the ‘redeemability’ characteristic of money leaves the model begging too many questions about why money is valued at all. And I think the doubts we can tell about such a model connect to practical concerns we might have about the real world, that such drops would completely undermine the monetary system, not save it.
This tale has a cross-over with the early part of the story of quantitative easing in the UK, in which I was a junior civil servant writing confusing memos not that dissimilar to this blog post. It became clear to me and others that most if not all of those who were deciding at that time didn’t know about the redeemability assumption explicit in the models. (I’d gone a large part of my time at the Bank not grasping it myself, and had to be talked through it by cleverer colleagues). They scoffed at it when it was spelled out to them. Their view was: well, the real world isn’t like that, so pumping all this money out there via QE will work. This is why the early communication on QE, and even now the educational material the BoE produces on the topic, stressed the monetary ‘injection’, and why there was no sweating about what assets to buy or that the MPC itself should decide this. If the only point was to get money out there, buying the assets was just a detail that functionaries in the BoE could carry out, not a policy decision. Only later, as evidence mounted that money expansion was immaterial, and buying duration in gilts was highly material, did the language, and eventually the policy behind QE, change.
Once again you might interject and say: Tony, they were right, redeemability was false, wasn’t it? Well, who knows. But whether they really understood this or not, the BoE ditched redeemability in their hearts, but still kept on working with a model that had it, but yet had to be then doctored with judgement to superimpose a stimulus they thought money would give that wasn’t there in the model itself. (There was none of the internal consistency, however questionable, of Buiter’s paper). So their forecasts were based on a fundamental, technical hotch-potch, assuming as it did that the stimulus would work through just like any other, when interest rates were away from the zero bound. But actually, in truth, the BoE were flying completely blind, and all the paraphanalia and whizz-banginess of the DSGE based forecast, which might have comforted some, was [actually, still is], on this account, something of a smoke and mirrors show.