Creating money without an accompanying asset – the Rowe – Keen discussion on bank assets

Nick Rowe has an interesting thread with a number of good contributions about Steve Keen’s Theory of Effective demand.

Here’s what I think Steve Keen is maybe trying to say:

Aggregate planned nominal expenditure equals aggregate expected nominal income plus amount of new money created by the banking system minus increase in the stock of money demanded. (All four terms in that equation have the units dollars per month, and all are referring to the same month, or whatever.)

And let’s assume that people actually realise their planned expenditures, which is a reasonable assumption for an economy where goods and productive resources are in excess supply, so that aggregate planned nominal expenditure equals aggregate actual nominal expenditure. And let’s recognise that aggregate actual nominal expenditure is the same as actual nominal income, by accounting identity. So the original equation now becomes:

Aggregate actual nominal income equals aggregate expected nominal income plus amount of new money created by the banking system minus increase in the stock of money demanded.

Nothing in the above violates any national income accounting identity.

Only last year he said he didn’t get it, but clearly has had second thoughts.

Much of this lack of understanding comes from confusion about the tools used. Those used to thinking in terms of ex-post accounting identities have struggled with whether there is an exposte- ex ante discrepancy. Those used to fixed periods and thinking of money as a stock struggle with how to conceptualise income and spending as instantaneous time flows. Those used to the GDP definition of income as deriving from value added struggle with how net purchases and sales of assets affect income. Those used to the loanable funds approach struggle with how demand can be affected by debt which they simply see as a transfer of existing income. Thankfully most but not all) of these issues have been tackled through Keens adoption of a double entry approach and its mathematisation with the aid of the Fields Institute. Most of the comments dealt with such issues of clarification and misunderstanding and I wont repeat them here, you can read the original thread and comments. Its us small usual bunch of monetary theory suspects im afraid, we could all fit in a telephone box.

Keen responded on Nick’s Blog

you’ve done a very good job of providing a Rosetta Stone between standard Neoclassical macroeconomics, and the perspective on endogenous money macroeconomics that I put forward

and then more fully on his own that

This is the first concerted (and very accurate) attempt to put my endogenous money approach in a form that Neoclassically trained economists can understand. This could be the start of a real dialogue in economics.

A quick note on Nick’s approach then ill look at an unresolved issue or two from the blog debate.

Keen has traditionally tackled the issue from actual realised income then placed back in the monetary circuit i.e. not saved (hoarded). Nick reformulates this in terms of the language of intertemporal optimisation beloved of neoclassical economics.

To give a very simple example lets say someone earned a disposable income of 10,000 dollars a month and they save 5,000 a month totalling 100,000 over time towards deposit on a house requiring a 500,000 loan. In previous months there actual nominal income 5,000 dollars equals expected nominal income (10,000 dollars) minus ‘increase in the stock of money demanded’ – savings in plain English – (5,000 dollars). In the month they get a loan there realized expenditure is 10,000 dollars minus, plus the 100,000 loan plus 10,000 disavings.

Keen has never explicitly covered the treatment of savings in this manner but I have long felt, and said a number of times on this blog, that his logic implies such a treatment. Indeed a number of historical predecessors of Keens approach to the circuit such as Norton and Johannssen.

Outstanding issues though concern what happens when a bank credits money without an accompanying financial asset and the difference between banks and nonfinancial institutions.

Rowe on this

Steve…If I sell my computer to my bank, the money supply expands. If I then buy that computer back from my bank, the money supply contracts.

Similarly: If I sell my IOU to the bank (if i take out a loan), the money supply expands. If I then buy that IOU back from the bank (if I repay the loan), the money supply contracts…

What’s special about banks is not what they buy with the money they create, but that they create money. And thinking about banks as buying and selling computers or land can help us make that distinction more clearly.

Steve replies

No, in the first case the bank is making a purchase of a commodity from you that it has to source from the liabilities and equity side of its ledger–not the asset side. To do otherwise is to commit seignorage–to use its capability to produce the IOUs we all use for transactions for its own use. So when a bank buys goods from non-banks, it uses the funds it has legitimately earned from its business of lending, not by using its capacity to create money. So there is no change in the money supply in either case.

He expands on this in his own blog

ending by an S&L … transfers money from its bank account to the borrower’s account, and therefore does not alter the total amount of money in existence. Lending by a bank … increases both the bank’s assets and its liabilities and thus increases the amount of money in existence. If workers try to get out of money and into gold .., they reduce their bank accounts but increase those of the dealers from whom they buy the gold….The only action that can take money out of the banking system is a withdrawal of money as cash

Nick Edmonds on the other hand, expanding his point on his own blog, disagrees. He argues from T accounts and accounting identities that non-bank financial institutions lending does increase the amount of money in circulation but not the overall stock of money and so this can have similar affects, though ultimately the ability to do so will be limited by the amount of savings they can attract.

Rowe in summing up his approach

Steve: I think (maybe) the biggest difference between us (in this context!) is that you focus more on the asset side and I focus more on the liabilities side of banks’ balance sheets. Let me state my view, to see if this clarifies things.

What’s on the asset side matters for a bank’s solvency and liquidity if people want to redeem their money. This matters a lot for commercial banks, which promise to redeem their money at a fixed exchange rate for central bank money. It matters also for central banks that promise to redeem their monetary liabilities at a fixed exchange rate for gold or USD or some other good. … If there’s a risk of insolvency or a bank run, the size and composition of the asset side matters. But provided the assets are good enough so we can ignore those risks, it is only the liabilities side that matters in the money-creation process.

So in my view:

A bank buying an IOU

A bank buying a computer

A bank buying a meal at a restaurant for its staff to celebrate Christmas

A bank giving money to charity

are all the same, in terms of creating money, and their effects on the liabilities side, though they will have very different effects on the asset side.

Ok the way I think about it is to distinguish between clearing banks and bank like institutions, shadows banks (including mutual funds utilising fractional reserve lending), and none bank financial intuitions (S&Ls, full reserve savings banks and the like) as each has very different mechanisms to back the credits they issue.

Whats special about banks?

Lets say a bank buys a computer, to use Nick’s example, if the computer company has an account at the same bank they can simply credit money in their account. If they don’t they can simply issue a cheque and create the money to redeem it, the end result is the same. The bank has increased its liabilities without creating an accompanying asset, the same as giving the money to charity. The key restraint on them doing so however is their balance sheet. A bank cannot issue money to buy the whole world’s supply of computers, they are restrained in doing so by the right hand side of their balance sheet, their capital, their equity, retained earnings and assets. Keen has tried to develop a theoretical approach whereby bank’s ability to lend is constrained by their charter value, the intangible asset that banks trade on from being a bank able to credit money and engage in fractional reserve lending/maturity transformation. I have tried to expand this into a mathematical model that takes account of banks capital and reserve ratios. I term this the lending power approach reviving a term from early endogenous banking theory.

Banks cannot expand their lending books without cost, and neither can they expend their asset of lending power on none loans (as in Nicks example) without opportunity cost. A bank forgoes potential profit on a loan in favour of a diversified portfolio which includes assets with a rate of return which may be higher than loans. This is the approach to bank portfolios which the later Toblin explored in several papers.

If a bank credits an account or redeems a cheque the Central Bank will accommodate that monetary demand. If they do so in terms of a loan then the creation of money will be cancelled out over the term of the loan by debt repayments and the bank as interest as profit. This interest payment is simply a transfer payment from the non-bank sector to the bank sector. If it credits a liability without creating an asset then it must run down an existing asset to stay solvent, which must come from pre-existing savings and profits. So the effect on the liabilities side is the same (Nick is right about this) but Nick neglects the constraints on banks ability to do so and that they can only do so through dissaving, which is already included in this framework.

Similarly Nick Edmonds argument is another case of a credit without an asset creation – as only banks and shadow banks can do so (the only difference between banks and shadow banks is that the central bank acts as lender of last resort for the former only, and clearance of final payments occurs through banks only), which also involves dissaving. Lets give an example. Lets say savings and loans typically create x billion of loans over a period and this is cancelled out by x billion of savings. Lets say interest rates change and there is a net increase of lending and net dissaving. Then from the formula above we have a net change in debt and a net dissaving, which cancel each other out.

Therefore only bank and shadow bank lending has a net macroeconomic effect.

10 thoughts on “Creating money without an accompanying asset – the Rowe – Keen discussion on bank assets

  1. You are potentially opening up a can of worms with your non-bank lending analysis. Non-bank lenders use warehouse money or white label products. If the funding is supplied by a bank it will increase the money supply as the warehouse lender (a bank) will create the funds, however a non-bank warehouse funder such as a superannuation fund won’t create any extra money supply.

    I think that Keen has it nailed.

    • There is a credit and a debt, a debit from the banks assets and a credit to the liabilities to those redeeming the payment for the computer. No conflict with double entry. The difference with a loan is it does not create a financial asset, rather it purchases a physical one. Both kind of assets however may generate future income streams.

  2. You are wrong in the example of a bank buying a computer or a restaurant or whatever, a bank can only create money when a loan is created at the same time. The loan is the banks asset, not a computer or any consumer item. Double entry bookkeeping must have a debit and a credit, but the asset must be a loan.

    Again Keen is correct on that issue.

  3. Andrew: “Only last year he [Nick Rowe] said he didn’t get it, but clearly has had second thoughts.”

    This is what I said in April 2012 (in the first comment in JW Mason’s post you linked to):

    “JW: I couldn’t make head nor tail of Steve Keen’s equation 1.1 either.

    Maybe, just maybe, somewhere deep down, what he is trying to say is something like the Wicksellian hot potato process. If the central bank cuts the rate of interest, planned expenditure exceeds expected income, and the stock of money exceeds the desired stock of money.”

    What I did in my post last week was flesh out that “Maybe, just maybe,…” hunch. And, from Steve Keen’s favourable response, it seems that my hunch was on the right sort of track.

    On your other point: yes, I have set aside the question of under what conditions banks would be willing and able to increase the supply of their monetary liabilities. I’m just looking at the consequences of their doing so, not the causes of their doing so.

    peter fraser: let’s take an extreme example: are you saying it is logically impossible for a bank to give money to charity, because this would violate some sort of accounting convention?

    • Nick thanks it was that comment to JW Mason I had in mind.

      A good example of banks creating a credit without an accompanying liability of the same NPV, of banks creating money and destroying less through debt amortization. Its when banks practice ‘forbearance’ as an alternative to default when the collateral backing the loan is worth less than the loan.

      Of course banks are in the M-C-M’ business, so if banks ran down an asset faster than building up a new asset from debt repayment, as they would do if they created money to give to charity, then the tier I capital requirements from the additional lending would rise but the rate of profit would fall so the bank would not be able to earn the capital from retained profits or equity. It would not be sustainable. Of course banks could create credits to purchase assets which would have a competitive rate of return, So I see no real difference from a bank using its balance sheet to purchase assets or to create them (debt contracts). Of course if all banks did was to purchase assets and not fund production we have problems. So by this reasoning banks are ‘special’ in being able to create net units of account, but this becomes problematic if the funding of assets from this credit (either through debt creation or direct purchase by banks of assets) produces assets price inflation eroding the income of debtors to the extent that they can no longer fund their debt repayments.

  4. Andrew,

    A nice post, but I’d like to clarify a few points if I may, particularly as they relate the post of mine that you referenced.

    First, though, a note on discrete vs. instantaneous analysis. I don’t think there is any fundamental difference in the results you get between the two. I usually (but not always) use discrete analysis, because solving the models I want to use is intractable analytically, so I have to use simulations which require discrete time. It should always be the case that the more you shorten the time period, the more it will approximate the continuous result.

    It is also worth noting that in reality economics does not take place in continuous time. It is a collection of discrete transactions that only approximate to the theory. The more we shorten our time periods, the more random and incoherent the behaviour appears. It is only by looking aggregating transactions observe a period, that patterns start to emerge.

    In my model, non-bank lending does indeed increase the amount of money in circulation, but it also increases the stock of money. For the purposes of the post, I have defined the money supply as the balance held in checking accounts, not the total of non-capital claims on banks. This was deliberate, given that the post was intended to focus on what determines the quantity of the medium of exchange. When non-banks lend in my model, the money supply is increased by households running down their non-money claims on banks. The point I wanted to make in that post was that payments money is not primarily determined by the amount of bank loans.

    If you are interested, I have expanded on this in my subsequent post, with some simulations and equation listings:

    Any lending involves asset creation. If I lend you £100, the amount of my assets is the same. I had £100 in cash, now I have £100 in loan to you. Your assets, on the other hand, have gone up by the £100 cash. Put another way, everyone’s debt is someone else’s asset. If debt goes up, assets must go up. On this point, it doesn’t matter whether the lender is a bank, a shadow bank or the local loan shark.

    None of this means that I think bank lending and non-bank lending are equivalent. There are important differences as I’ve illustrated in my subsequent post. It’s just that I think it is quite wrong to believe that only one class of debt matters.

  5. “If it credits a liability without creating an asset then it must run down an existing asset to stay solvent, which must come from pre-existing savings and profits.”

    I presume you meant to say it must run down an existing liability. If a bank makes a payment without acquiring an asset then it reduces the balance of its internal accounts recording its retained earnings, which are on the liabilities/equity side of its balance sheet, and credits the account of the payee, increasing its deposit liabilities to restore balance with its unchanged assets. If the payee is not an account holder, then it must give up assets to balance its reduce equity by transferring central bank reserves in settlement to the payee’s bank, which will itself credit the payee’s account, increasing its own liabilities to match its increased assets.

    If the bank acquires a computer or other asset from a non-bank entity then it simply increases its liabilities to match by crediting the supplier’s account. Again, if the supplier is not an account holder then matching assets will instead be transferred in settlement to the supplier’s bank which will in turn increase its own deposit liabilities by crediting the supplier’s account.

    Whichever way the transactions go, whenever a bank makes a payment to a non-bank entity, deposit liabilities increase The official measure of the money supply increases.

  6. In my own post with banks and non-banks, the two alternatives for new loans are:

    1. Bank. Debit loan (increase asset), credit borrower’s deposit (increase liability).

    2. Non-bank. Debit loan (increase asset), credit cash-at-bank (decrease asset).

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s