More from physics, mea culpa, bond behaviour - and two suggestions

Lots in this Post. Firstly, maybe there’s another analogy from physics, that can be applied to accounting.

The conventional practice that Steve has described, of always creating financial assets and liabilities in matching pairs, looks quite hunky-dory to me when money always arrives from outside the account. But as far as I can see, that practice goes logically wrong when money is created inside the account, as by fiat. It incorrectly says a debt is due where none exists. I’m reminded of how the rules of physics must be altered somewhat when the system described is rotating. Does fiat money creation, seen through this analogy, put accounting into a spin?

I think two matching assets, of opposite sign, must recorded without the recording of a net liability to another party, when money (or funds, meaning bonds) are created by fiat. But that would violate accounting practice, I think. Earlier, I’d thought that another kind of liability - a self-liability - would do the job of allowing conventional accounting to correctly describe fiat money creation. Mea culpa - I was wrong, I now admit. I’m grateful to @Chris Rimmer and @Kevin Carney and other Skool members for pointing out my mistake. I had misled myself by using the case of a physical liability - carrying a heavy raincoat on a hot, dry day, because intense rain was wrongly forecast, means the coat is liability to carry - a self-liability of its owner.

However, financial liabilities are surely a special subset of the general meaning of the word, that specifically describe an obligation to pay the target of the liability. So a financial self-liability, matching an asset, is an obligation of the owner of the asset to pay it to the owner of the asset. This is a null operation - nobody can be in debt to themselves. So I now feel that there can be no such category of liability - a self-liability is a contradiction in terms, since it is discharged by doing nothing and losing nothing. Therefore, a self-liability is not a suitable vehicle for recording the negative component of fiat money creation. It has no net negative effect on financial liability, since it increases assets by the reduction in assets that occurs when the liability is enacted and discharged. Ooops - rethink required, no?

This situation seems to arise in Steve’s Ravel/Minsky description of State money creation by fiat. To cancel the Treasury’s overdraft, that results from spending more than tax revenue, a Treasury (T) bond is created by fiat, recorded in conventional terms as a matching asset-liability pair. The asset (+ value ) bond is asset-swapped with a bank for Central Bank (CB) reserves, thereby erasing the overdraft.

But surely a problem has arisen with this accounting, not arithmetically but logically I think. T’s accounts now seem to suggest it has fallen into debt to the bank holding the bond - not just for the bond’s interest but for its principal sum too. Is the T-bond a loan with an expiry date, which the receiving bank maturity transforms into current value? I think that question is answered when the bond reaches its expiry date and the interest flow then ceases. Is the bond ‘bought back’ by T?

If the bond was a loan, surely that would be how the loan is repaid. In that repayment, the money previously created would be destroyed. This would mean that States with their own currency do indeed borrow when they create money, in the same manner as citizens do when obtaining bank credit, and the National Debt is genuinely a debt the State owes to banks. But I understand that T-bonds are ‘bought back’ (wrong phrase, surely) by the CB. They are deposited into the CB by the owning bank, which receives reserves in return. Is this true?

That can only happen if the T-bond is not a loan of funds but a financial instrument of permanent net value - the backing for the fiat-created money with the same permanence. If so, what then of the bond-liability in T’s accounts? It surely represents a debt obligation, falling upon T’s shoulders, to undertake the action of payment of that amount to the bond holder - the CB. The ownership of this liability has put T into negative equity, but it has no asset with which to ever discharge the liability - it cannot ever pay the debt. How can T be liable to perform an act that is impossible to perform? Surely what has occurred, logically, is that T is holding not a liability but a negative asset, which is one half of a matching pair - the T-bond in the CB being the other half. I think it would reveal the truth of the situation, when fiat money is created, if a negative asset was recorded in T's accounts instead of a liability.

This need to violate accounting practice, in order make accounting logically meaningful as well as arithmetically correct - seems to be a special case, that applies only when money does not enter from outside but is created from within, as by fiat. I feel that discussions at Skool have tended to lead towards two possible kinds of corrective measure -

A) alter the meaning of a financial liability so that it has two cases. In the first case, it means debt - an obligation to undertake the action of payment. In the second, and more restricted, case it means no obligation to undertake an action of payment but instead expresses negative value that is free of debt. That second meaning seems to be contrary to the commonly accepted meaning the public appear to hold. Maybe I’m unjustly putting words into their mouths, but that what I’ve understood to be the effect of Challenge Skool comments from Chris and Kevin.

B) is a suggestion of my own, to retain the conventional practice of accounting in creating asset-liability pairs, while breaking with what might only be an accounting preference rather than a strict mandate. I suggest the use of negative numbers in both assets and liability entries. Taking the T-bond example, as above, it would mean creating the bond as asset and liability as before (positive numbers) and simultaneously creating another matching asset-liability pair, of the same magnitude but with opposite sign. So a negative bond asset and a negative bond liability get created at the same time as the positive versions. Rather than creating one matching pair, create two pairs with the second carrying opposite sign - specifically limited to the case when money appears without entering the accounts as a deposit from outside.

When the normal bond is asset-swapped for CB reserves, T still falls into the same amount of negative financial equity as before. But there are now two liabilities borne by T - the first is (in the usual meaning of financial liability) a debt of T to the bond-holding bank, while the second is the reverse - a debt of the bond-holding bank to T. So no net debt of T now exists. The negative equity needed to create the positive equity of the fiat money is now recorded by the negative bond asset, held by T. Yes, it’s a rather cranky and cumbersome procedure, but it fits the ‘asset-liability-pair’ practice and makes it work when money ‘enters’ the system from within not from without. It therefore does seem to express the truth of fiat money creation - that fiat money is created without debt by means of negative financial equity held by the money creator.

I wonder if @Stephen Keen might comment, to sort out the wheat from the chaff here?

0 comments