Quantitative Easing Redux

A couple of weeks back, I posted on the topic of “quantitative easing,” the policy of having the central bank aggressively purchase government (and possibly corporate) debt in the open market ostensibly to increase the money supply.

I argued that at best, quantitative easing was a pricing operation that worked at the margin by increasing demand for a given asset class. Since then, the Bank of Canada has more or less announced that it will engage in this policy, with details forthcoming in its April 23 Monetary Policy Report.

Now, Bill Mitchell — a fellow traveler in the Post-Keynesian world (from Australia) — has posted a more comprehensive analysis of quantitative easing which makes the point that the very term “quantitative easing” implies a misunderstanding of how money and monetary policy work in a modern economy.   I don’t want to pre-empt Dr. Mitchell because he does a bang-up job of making everything crystal clear, but I do want to emphasize four points that are absolutely crucial for understanding this issue:

Proposition 1: Banks create money every time they make a loan.  They worry about reserves (or settlement balances in Canada) later (again, see the excellent Baumol and Blinder textbook (canadian version by Marc Lavoie and Mario Secceraccia) for a detailed analysis).  Even the Bank of Canada acknowledges as much, when it notes: “Commercial banks and other financial institutions provide the greater part of assets used as money through loans made to individuals and businesses. In that sense, financial institutions are creating money.” Contrary to what you read almost everywhere in the media and from the industry itself, banks don’t “borrow first and then lend (so-called “on lending”).  They lend first, then if need be (and this is a settlement balance and liability management issue), borrow.

Proposition 2: The Bank of Canada — and most modern central banks operating in sovereign currency environments — are quantity takers and price setters.  That’s why they’ve stopped measuring esoteric versions of the money supply. That’s why they gave up on monetarism in the 1980s — they couldn’t make the definitions do what they said they should.  That’s why they call money supply “a residual.”  Incidentally, this proposition follows from, and is integral, to the first.

Proposition 3: If the Bank of Canada is contemplating “quantitative easing,” that means it is also contemplating a zero target for the overnight rate, as Bill makes clear.  I should have insisted on this point in the previous post.

Proposition 4: The financial system has not matched the Bank of Canada’s price declienes (i.e., lowering of interest rates) because of default risk and, to a limited extent, because their cost of borrowing on international markets for foreign currency– and perhaps as well, their liability management efforst — have gone up (and, incidentally, a big chunk of the increase in chartered bank lending — of which the Canadian Bankers Association is so proud — is due to lending in foreign currency to Canadians, presumably exporters who are struggling — see the Bank of Canada’s Weekly Financial Statistics for details).  The international interbank lending market is a mess, as any look at LIBOR reveals (although things have improved of late). Quantitative easing won’t change either fact except perhaps through some secondary effect on expectations.


  • Steve Keen did a nice job at explaining the difference between the standard neoclassical explanation and the Post-Keynesian explanation. Some nice graphs too.


  • Hey Arun,

    You are easily the best masked economist in Canada right now, and your comments are right on.

    I’ve been pushing the Laidler-eque version of quantitative easing, whereby the current deficits of the federal government are financed through the Bank of Canada, with dollars going into public hands through income transfers and public job creation, thereby stimulating the demand side. I think this would be a harmonious blend of monetary policy and fiscal policy.

    What do you think?

  • Fresh off the Globe’s website:

    “Taking their cue from Canada, the finance ministers of the 20 largest industrialized countries agreed that their “key priority” was fixing the banks so they can start lending again and lubricate the economy. Mr. Flaherty has been relentless in his warnings that economic recovery is impossible unless the banks are fixed, which means ridding them of their toxic assets.”



  • Hi Marc,
    Thanks for that — I like the handle! On the substantive point, I wholly agree with the proposition that as far as the crisis is concerned, policy needs to work on the demand side and income distribution and job creation are the way to go (in fact, I would advocate a guaranteed job program of the kind put forward by the folks at the Centre for Full Employment and Price Stability — http://www.cfeps.org/).

    My only point of disagreement would be a theoretical / institutional one. I take the view — espoused by Warren Mosler and other “chartalist” theorists — that every act of spending by a sovereign government (such as ours) with its own sovereign currency (i.e., no currency boards and/or no dollarization) entails money creation, or, in the Canadian case, an increase in settlement balances in the financial system (in a non-zero overnight rate environment, govt spending compels the Bank of Canada to neutralize the spending with a settlement balance drain but fortunately, we’re going to zero soon so that’s not an issue).

    To put matter succinctly, sovereign governments don’t need — and operationally, do not — “borrow” from the private sector to spend their own currency. I realize this sounds a bit strange but if you look at the actual institutional mechanisms this is exactly how it works — governments spend first and tax — or borrow — after to drain reserves from the system.

    Randall Wray (see CFEPS and google searches) and assorted fellow travelers have done some brilliant historical work that situates this argument in its proper historical context, namely that money historically is as much — if not more — a creature of the state than a spontaneous market response to a coordination problem (the traditional narrative). And from this perspective, money derives its ultimate value from the power of governments — or whatever authority who runs the show — to impose taxation. By the way, the guaranteed job proposal put forward by Wray and Mosler is designed to dovetail with this understanding of money and spending.

    I’ll close by noting that Warren Mosler has some really quite nice pieces that summarizes this perspective for the layman (go to his “Required Reading” section). For a more technical presentation, I strongly recommend anything by Scott Fullwiler — very lucid, technical and clear presentations. Check out his Journal of Economic Issues papers especially.

  • Travis,
    Read Keen’s piece — very good. You might also want to check out his book — Debunking Economics — nice, tightly argued critique from someone who understands the math better than most (a lot of of Keen’s academic work is related to modeling from a post-keynesian (non equilibrium) perspective).

  • Arun,

    It has been on my shelf since it was first published. His blog has been good for ameliorating the effectiveness of his prose.

    What do you think about his argument on the ratio of debt to income ? It would seem from that argument that essentially the way forward is a reduction of household debt to income. Right now it seems we are choosing some form of the austerity route which is long and painful. Governments can spend now and stabilize the economy, but then they will have to chase the tax dollars later suggesting that demand will be sucked up by public debt repayment. We could have a legislated haircut of outstanding consumer debt but that would destroy bank money and imply a reduced willingness to supply credit because even if banks lend first and search deposits later they must have some optimism that they can find those deposits, i.e., depositors are likely to be gun shy if they just took a 20% haircut. And I suppose we could hope that real wages actually increase (with the necessity of extending and increasing EI benefits) during a prolonged period of deflation in other prices thereby eroding the debt to income ratio (not to mention the erosion of financial profits). None of these options suggest a quick or painless turnaround.

  • Hi Travis,
    Good points. My understanding is that Keen approaches the debt issue largely from a Minskian perspective — which I would summarize as follows: private-sector debt accumulation (mostly due to business activity but clearly household behaviour also contributes) explains most boom and bust cycles.

    My understanding is that Minsky was skeptical of our ability to completely avoid this process but he did put a lot of emphasis on the need for government to step up to the plate and regulate, redistribute and otherwise save finance capitalism from itself.

    On your point about government having to suck up demand at some later point (if it bails things out), I’m not sure about that. I mean that literally — I’m not sure! That said, I would only note that Minsky did — in some places — advocate a guaranteed job program of the kind put forward by the CFEPS people. That seems to me a reasonable way out of this mess.

    On forgiving consumer debt ( a jubilee anyone?), I think your analysis is correct unless of course the feds step in and fill the gap, compensating the banks for any losses (not beyond the pale in the current environment).

    As for real wages, that again points me to an ELR — employer of last resort — type policy which would have positive real-wage effects by creating a real-wage floor that the private sector would have to respect (for the most part — some exceptions can be contemplated).

  • “Proposition 1: Banks create money every time they make a loan. They worry about reserves (or settlement balances in Canada) later (again, see the excellent Baumol and Blinder textbook (canadian version by Marc Lavoie and Mario Secceraccia) for a detailed analysis).”

    To be absolutely clear:

    Banks create credit first then wait for that credit to settle back into the banks where it then balances out the liability and asset columns

  • Pat,
    Good question. I was in fact worried that there may have been some confusion about this issue.

    By double-entry book-keeping practices, every loan (asset to the bank) has, as its counterparty, a deposit (liability to the bank). From the borrower’s perspective, the loan is the liability and the deposit is the asset.

    The resulting deposits can flow to other banks (when customers write checks to customers at other banks) and lead to settlement balance deficits, which is why banks engage in liability management, i.e., try to match the terms on their loans with their liabilities — they try to sell us, in other words, 5-year GICs and such so that our deposits are parked for a period roughly equal to (say) a 5-year mortgage. Canadian banks are fortunate in that because of their size and national scope, a lot of their clients probably write checks to other clients at the bank, minimizing the scope of deposit drain.

    In the US, deposit drain is a more immediate concern because most banks are small in size and/or regional in nature. Some US banks tend to have excess deposits (relative to their daily clearing needs) and auction these off to chronic deficit banks (like the NY banks who engage in a lot of lending).

    I think it really helps to think of each bank as creating its own branded version of money (with every act of loan creation). Problems arise when bank customers try to interact with customers from other banks.

    Hope that helps clarify my position, which is essentially the Post Keynesian position BTW.

  • Thanks for the post and the link, both were interesting and I’m generally in agreement. Like Travis, I generally see reducing the (public+private) debt load as the key to putting the macroeconomy back in order.

    A couple of questions:

    Might the purpose of the quantitative easing be to support asset values and encourage lending in this manner rather than by simply increasing reserves? (I’m not saying that’s a good idea, mind you).

    Also, Bill comments that banks are constrained by their capital, not their reserves, but won’t buying up the bank’s assets improve their capital ratios by substituting cash (capital) for assets (which require capital kept against them) as well as increasing the mark to market value of the remaining (unsold) assets?

    Unless I’m missing something (which wouldn’t surprise me) a QE policy could raise bank’s capital levels and the higher a bank’s capital level, the more willing to lend it is, all else equal.

  • Arun Dubois

    Thats probably a good way to think about the credit each bank creates. After all, before the single currency was created each bank printed its own notes to represent its own credit.

    I remember reading about people with lots of “money” who couldn’t buy anything because the bank their notes were from gained a bad repulation and nobody would accept their notes. Of course when they tried to withdraw their commodity money that bank had nowhere near enough to cover all the notes they had created.

    When I visited Hong Kong in the early 80’s I was floored to find out they had two separate currencies. Growing up here I had no concept of privately created money supplies (its that well hidden in our systems) and I couldn’t figure out why the government would print and coin to different versions.

    Our countries had government sanction a single “brand” (Bank of Canada Note) and all the banks are now OEMs creating their own credit under the BoC brand.

  • Declan

    The BoC collecting capital only increases the BoC’s capital reserves and that allows it to lend more to regular banks. Regular banks have to be willing to borrow from the BoC to make loans of their own but if they aren’t willing to loan then they probably aren’t willing to borrow from the BoC to do it.

    The amount of credit banks are allowed to create is based on their capital. The BIS sets allowable capital to deposit ratios. After Mulroney(or was it Martin?) eliminated reserve requirements and switched to capital ratios that ratio was roughly 20:1 deposits (liabilities) to capital (depends on bank size etc). I think under Basel II that was increased to 40:1 but I’ve not researched Basel and I’m going by hear-say mostly.

    Bank Capital is government bonds, interest profits kept by the bank, and any real property.

    Assets and liabilities (loans and deposits) just have to balance out over the settlement period.

    The US is slightly different in that they still have a tiny reserve requirement on demand deposits of 4% – but only if the deposit exceeds a few millions of dollars. The reserve requirement is a joke really since all the rest of people’s deposits have no reserve requirement. If you have no reserve requirement on a penny then you could loan out trillions of dollars based on that penny.

    Anyway, if a retail bank sold a government bond to the BoC it wouldn’t change its capital or asset position – that cash would be capital, not an asset – so they’d just swap an interest bearing bond with cash.

    If the cash was lent out that would reduce the bank’s capital position and make the asset/liability larger but still balanced. The reduced capital position could (and likely would) put them over the BIS capital to liability ratio.

  • Pat, the deposits are irrelevant for capital purposes under both Basel I and II, it is assets (risk weighted assets, specifically) that determine the capital requirement.

    Selling a government bond would have smaller capital implications because the amount of capital kept against government guaranteed assets is minimal.

  • I should correct my first comment slightly:

    The banks ability / willingness to lend is constrained by the ratio of its risk weighted assets to its capital (effectively, its equity).

    QE can improve this ratio in 3 ways:

    1) By giving the bank an asset (cash) with a 0 risk weight instead of one with a positive risk weight. (note, this effect will be small if the BoC purchases govt bonds that have a low risk weight, the lower quality asset the BoC buys, the bigger the impact) (note also that my first comment saying that cash *was* capital, was mistaken).

    2) By allowing banks to sell assets for a higher price than otherwise making for a more positive equity impact of the sale

    3) By allowing banks to sell assets for a higher price than otherwise allowing them to mark the assets they don’t sell at a higher price (assuming they are using mark to market accounting – if not, this wouldn’t apply).

    Reserves and deposits are not really relevant, as far as I can tell.

Leave a Reply

Your email address will not be published. Required fields are marked *