This paper conflates two quite different things, namely full reserve banking and debt jubilee, which causes a fair degree of confusion and I feel it would have been sensible to have separated them. It was also very evident that the authors were far more comfortable with economic modelling than they were with accounting and financial flows, despite their claim that one of them had actually done loan accounting for a bank. The accounting models in particular are in my view seriously flawed, and insufficient consideration has been given to the impact of the suggested changes on financial flows and the role of banks. I don't propose to evaluate the authors' macroeconomic assessments in this post: I shall confine myself to an analysis of the accounting models and the effect on the financial system.
Pages 4-8 of the Introduction are concerned with the balance sheet impact of the move to full reserve banking and the debt jubilee, and they are supported by three accounting models (Figures 1, 2, and 3) at the end of the report. I'm reproducing the first two models here for convenience, and I have extracted the relevant bits of the paper here.
Model 1: changes to aggregate private bank balance sheets (click here for larger version)
Model 2: changes to government/central bank consolidated balance sheet (click here for larger version)
Even with the models it took me quite some time to understand how the accounting worked, basically because it doesn't. However, before I take a hatchet to the models, I need to note and briefly discuss some assumptions that the IMF writers have made but failed to document.
Firstly, they do not distinguish between central bank and government. This conflicts with the IMF's long-standing preference for central banks to be independent of government. Is the IMF about to change its views on this? The writers treat central bank accounts as consolidated with government accounts. This is reasonable accounting practice where the central bank is fully state-owned, as in the UK: indeed the UK Government produces "whole government accounts" (WGA) that include the central bank. But this paper is written from a US perspective. The Feds are currently privately-owned. They cannot be consolidated with government. So, undocumented assumption #1 is - all central banks that are not currently nationalised will be.
The effect of consolidating central bank and government accounts is to show government debt and money in circulation on the liabilities side of the same balance sheet. The IMF doesn't show this, but we know from the UK's WGA that government debt bought by the central bank is eliminated from the consolidated accounts, replacing it with the central bank money issued to purchase it. In effect that debt becomes money. As I've noted before, the problem is that although the accounting consolidation "eliminates" the debt, the debt instruments themselves still exist in reality. As government debt cancellation forms part of this proposal, this problem arises in the IMF's model. More seriously, it also arises with respect to money in relation to the cancellation of private debt (debt jubilee), as I shall discuss later.
Unfortunately the IMF's view of the capital structure of private banks is not so realistic. The IMF's model shows private banks' liabilities as consisting solely of deposits and equity. There is no mention of bonds or interbank borrowings: yet these days banks are financed to a considerable extent by bond issuance and interbank borrowing. Few large retail banks have a deposit base of a similar size to their total loans, but this is what the IMF's model shows. It is over-simplified and unrealistic.
This is serious, because the accounting for the proposed "debt jubilee" depends on loans being balanced by deposits. The writers seem to assume that "deposits" includes bonds. But what about interbank and repo balances? Are they "deposits" too? Whatever, it is clear that undocumented assumption #2 is - private banks' debt consists entirely of customer deposits.
There is a similar problem on the other side of the balance sheet. The IMF writers assume that bank assets consist only of various forms of lending to domestic corporates and households, plus a small quantity of government bonds. Oh, if only that were true! The fact is that banks are major investors in non-government bonds, equities and derivative instruments, and they lend to each other and to non-bank financial institutions. And many of their assets are cross-border and denominated in foreign currency. Undocumented (and totally wrong) assumption #3 is - private banks lend only to domestic households and corporates.
So, having made three assumptions that are both wrong and undocumented, the writers then go on to explain the accounting for full reserve banking and the associated debt jubilee. The key components are as follows.
1. Banks must hold reserves in the form of government-issued money to an amount equal to the amount of customer deposits they hold.
As customer deposits are a moving target, I assume this means that the central bank will have to provide reserves on a daily basis as at present, though the paper doesn't say so. However, the central banking accounting for reserve creation is different in the IMF's model. There are no open market operations or collateralised lending. Instead the writers create a "financial asset" called "Treasury Credit" to balance the reserves created to back customer deposits. Effectively the CB will lend money to banks (unsecured) every time a customer deposits money, and the balancing asset will simply be a debt liability of the receiving bank. This is fractional reserve banking as we know and love it - a bank creating a deposit ex nihilo balanced by a loan account - but done by the central bank in relation to private banks on a simply massive scale. The result would be that banks would suddenly have enormous debt liabilities with regard to the Government/CB, and Government would have a balance sheet vastly expanded by the inclusion of the total of all domestic customer deposits (and possibly wholesale ones as well) in the form of equity. Figure 2 in the IMF paper shows this.
This is not so far removed from the "narrow banking" preferred by the Bank of England's Mervyn King. The difference is that reserve backing for deposits is provided by money, not by government debt. There are unresolved issues with the IMF writers' proposal because they have over-simplified private bank balance sheets. But otherwise, the accounting makes sense provided that assumption #1 holds.
If bank bonds were included in "deposits", then they would become risk-free assets equivalent to Government debt, or alternatively they might be redeemed. The paper doesn't say, but I assume that repo and interbank borrowings would be redeemed rather than backed, since banks would no longer be allowed to obtain funds from each other. This brings me to component two - the changes to bank funding.
2. Banks will not be able to use customer deposits (or any other sort of private borrowing) to fund lending.
All lending must be funded either from own equity or borrowing from government/CB. This last is particularly bizarre as it means the bank would have large amounts of money sitting around doing nothing, yet it must borrow EVEN MORE money in order to lend.
At this point I get irritated with the writers' unrelenting balance sheet focus. They seem to have completely forgotten about the profit & loss account and, indeed, the purpose of private banks. The sole aim of private banks is to make money, principally through lending. And the sole point of taking deposits is to fund lending. If deposits couldn't be used to fund lending, why on earth would banks want them? Deposits that couldn't be productively used, and that would moreover COST the bank money because of the 0.49% proposed interest charge on reserves required to back them, would not be commercially viable. Banks would inevitably look for some way of making money from depositors, probably in the form of safe deposit fees and management charges. The paper suggests that the real interest rate on deposits would be zero. I beg to differ. I think the real interest rate would be negative. Banks would have to be able to make money from them somehow, especially given the severe curtailment of lending and its low margin under this proposal.
Not only would banks not want to accept deposits, savers wouldn't want to make them. The model assumes a real interest rate of 1.07% on government bonds. Why on earth would savers (except very short-term ones) put money in a deposit account at 0% or less when they could buy government bonds at 1.07%? It seems likely that the effect of the rate inversion (deposit rate < reserve rate < government bond rate) would be to restrict bank deposits to safe deposits and checking accounts only. Time deposits and "at risk" sight deposits would become a thing of the past - banks would have no need to offer them since they would not need them for funding, and savers would have no reason to want them since there would be a plentiful supply of government bonds at a much better interest rate.
So this rule would be likely to end much bank deposit-taking. It would also end shadow banking, which is essentially a chain of deposits repeatedly on-lent against safe collateral. But from a lending perspective, the picture is much less clear. This rule alone would not restrict commercial lending, but there is no explanation of how reserve interest would be paid, or how money supply inflation would be managed
Consider a bank making its first loan. The bank must obtain funds from the central bank in advance of the loan being agreed (since the money supply inflates when the book entries for the loan are made, and it is that inflation the writers are trying to eliminate). Those funds are loaned to the bank at the reserve rate. Because this is its first loan, the bank has no interest income, so it must either pay the interest from its own capital or it must also borrow from the central bank the money required to pay the interest. If the latter, then this funding model is a Ponzi scheme.
As at present, the book entries for the loan create a deposit, but this time it is 100% backed by new reserves already provided by the central bank. So the loan STILL inflates the money supply - it's just that the central bank creates the new money rather than the lending bank creating it. Is the decision whether or not to grant the loan going to rest with the central bank - or is the central bank just going to do what the lender wants without question? If the latter, in what way is this form of money creation an improvement on the present situation? If the former, in what way is the central bank better placed to decide on the merits of a loan application than the lender?
When a deposit is drawn, the person or business receiving the money deposits it in another bank. Will reserves automatically transfer from lending to receiving bank, even for smaller banks and shadow banks that don't currently use central bank settlement accounts? If so, then the deposit would remain 100% backed by reserves when it moves from one bank to another. If not, then reserve imbalances could quickly build up, leading to deposits becoming "at risk" in net lenders. I foresee quite a lot of systems enhancements to extend central bank settlement facilities to those players that don't currently have access to central bank settlement facilities.
3. Existing household and short-term corporate loans will be bought out by government/CB (debt jubilee).
This is where it all goes very, very wrong. The IMF writers, gleefully looking at a Government/CB balance sheet vastly inflated with huge amounts of equity, wish to take the opportunity to get rid of most corporate and household debt. So they buy an asset - customer loans from private banks - and write it off against equity. But in so doing they have forgotten that the equity is not simply shareholders' funds and retained earnings, as it is in a corporation. It is real money that is used in the real economy to buy real goods and pay real wages. If the equity on the Government's balance sheet were reduced the real money in circulation wouldn't reduce, and the result would be an enormous imbalance between the book records of money issued and the money actually in circulation (nearly all of which would be recorded in the books of private banks, and on which they would be paying interest). Figures 1 and 2 (third column) show this clearly. In effect, the Government would have issued a very large amount of interest-bearing money that it had not accounted for. I think they call this accounting fraud.
If they actually tried to balance the public records with private bank reserves, the result could be even worse. They would have to instruct all private banks to write down reserves by the value of the loans written off. As reserves would be the backing for deposits, the value of deposits would also have to reduce by the same amount. That would be a MASSIVE hit to depositor value. Effectively they would have trashed the currency (since it is the value of money in circulation that is written down) and people would have lost virtually all their savings.
The alternative would be for the private debt to remain on the central bank's balance sheet, where it could be segregated and treated as non-performing. This would have no impact on reserves, but would reduce the Treasury Credit. In effect, the central bank would be accepting private debt assets as settlement of the Treasury Credit (loan) imposed on banks in the move to full reserve banking. However, it can't be assumed that the Treasury Credit would match the total value of household debt and short-term corporate debt. If there was a shortfall for a particular bank (this is quite likely in large retail banks reliant on wholesale funding), would the central bank force the private bank to accept more reserves than it needs? As the writers envisage the repayment of debt being done by the borrowers themselves using money paid by the central bank into segregated customer accounts that could only be used to repay debt, it seems the answer is yes, it would. But the bank would have to pay interest on those reserves. The only way it could make money to pay the interest on those excess reserves is to lend more. Which brings us to component number four - the future purpose of bank lending.
4. In future, banks will be allowed to lend only for investment projects.
This does of course raise the question of how future lending needs other than investment projects would be met. Mortgage lending, for example. No provision exists in the model for banks to lend to FUTURE households to enable them to buy homes. Are the authors assuming that ending domestic mortgage lending will force a house price crash of a sufficient size to enable the majority of future households to buy houses without borrowing? Or are they assuming that people in future will rent rather than buying - and if so, how will landlords finance the purchase of property for letting? Or are they assuming that Government agencies (Fannie & Freddie) will lend directly to households, rather than indirectly via mortgage originators as at present? They don't seem to have thought about this at all.
If banks were restricted to lending for investment projects only, lending volumes would of course be seriously reduced. And furthermore, banks would make far less money from their remaining lending, since spreads would be tiny because the corporate debt write-off is assumed to reduce risk. The model assumes that corporate borrowers could alternatively obtain finance for investment from the capital markets at close to the risk-free rate, and that therefore bank lending rates could not exceed that. If that is the case, then bank lending is likely to be unprofitable given the cost of reserves required to fund it.
So, to sum up this rather long and detailed analysis: as far as I can see, in this model it is virtually impossible for private banks to be profitable. In which case they would soon cease to exist, and government would be forced to create state banking facilities to replace them. The question is, why did the authors stop short of recommending full nationalisation of the banking system, since that is the only way the model could work in the longer term? The authors think that private banks funding long-term investment projects is a Good Thing, but they offer no explanation for this belief. Could it be that they have retained private banks in their model because recommending a move to a wholly state-owned system would not be taken seriously by most economists and politicians? After all, state ownership of banking has been a far-left wet dream for a very long time. I'm not a fan of a wholly state-owned banking system, but it would surely be better than a system in which supposedly private banks are prevented from making commercial lending decisions, forced to borrow (and pay interest on) money they don't need and unable to make profits.
This proposal is poorly thought through. Or maybe it isn't. Maybe the real purpose of this proposal is to force the death of commercial banking, and in particular, shadow banking. For that is what it would achieve if it were implemented.