Believe it or not I’m still getting a lot of questions about how QE works,
so I’ve reorganized the discussion some:

First, recognize that, in fact, reserves, functionally, are nothing more than 1 day t bills.

And, for all practical purposes, the difference between issuing 1 day t bills and 3 month t bills is inconsequential.

So since currently the shortest thing the Treasury issues are 3 mo bills,
I can say that:

QE- the Fed buying longer term treasury securities- is functionally identical for the economy to the Treasury having issued 3 month t bills instead of those longer term securities the Fed bought.

Now the more tricky part.

The yields on the approx. $13 trillion of various Treasury securities and reserve balances continuously gravitate towards what are called indifference levels.

That means that for any given composition of reserve balances at the Fed and Treasury securities (also Fed accounts), there is a term structure of interest rates that adjusts to investor preferences at any given time.

So, for example, with govt providing investors with the combination of $2 trillion in reserves and $11 trillion in various Treasury securities, the yield curve will reflect investor preferences given the current circumstances.

That means if investors expect Fed rate hikes, the front end of the curve would steepen accordingly. And if instead they expect 0 rates for a considerable period of time, the curve would flatten for the first few years to reflect that.

If the Fed then buys another $1 trillion of securities, reserves go to $3 trillion and there are $10 trillion longer term Treasury securities outstanding.

And to actually purchase those reserves, the Fed would have to drive the term structure of rates to levels where investors voluntarily are indifferent with that mix of offerings, given all the other current conditions.
The Fed doesn’t force anyone to sell anything.
It just offers to buy at prices (interest rates) that adjust to where people want to sell at those prices.

So even if the Fed owned a total of $10 trillion of securities, and there were only $3 trillion left outstanding for investors, if investors believed the Fed was going to hike rates by 3%, for example, the term structure of rates on Treasury securities would reflect that.

What I’m trying to say is that QE does not mean rates will actually go down. The yield curve is still a function of investor expectations.

But the yield curve is also a function of ‘technicals.’
This means the quantity of 30 year securities offered for sale, for example, can alter the yield of that sector more than it alters the yields of the other sectors.

This is because, in general, there tends to be fewer ‘natural’ buyers of 30 year securities than 3 month bills.
For most of us, we are a lot more cautious about investing for 30 years at a fixed rate than for 3 months at a fixed rate.
And it takes relative large moves in 30 year rates to cause those investors to shift our preferences to either buy them if govt wants to issue more, or sell them if the Fed wants to buy them back.

On the other hand, there are pension funds who ‘automatically’ buy 30 year securities regardless of yield because they are matching the purchases to 30 year liabilities.

So altogether, the yield curve is function of both investor expectations for interest rates and the ‘technicals’ of supply and demand (desires by issuers and investors).

And while there might be no amount of 3 month bills the Treasury could issue that would materially drive up 3 month t bill rates, relatively small amounts of 30 year bonds do alter the yields of 30 year securities. Insiders would say the 30 year market is a lot ‘thinner’ than the 3 month market.

So what is QE?

QE is nothing more than the govt altering the mix of investments offered to investors.

The Fed buying longer term securities reduces the amount of longer term securities and increases the amount of reserves (one day securities)

Interest rates, as always, continuously gravitate to reflect current investor expectations of future Fed rate changes and current ‘technicals’ of supply and demand.

QE changes the technicals, and possibly expectations, and results in a yield curve that reflects those current conditions.

So all QE does is alter the term structure rates, as investors express preferences for the term structure of interest rates, given the securities and reserves of the varying maturities offered by the Fed and Treasury and all the current conditions.

That brings us back to the question of what QE means for the economy, inflation, value of the currency, etc.

Which comes down to the question of what the term structure of rates means for the economy, inflation, the value of the currency, etc.

QE is nothing more than a tool for changing interest rates by adjusting the available supply of securities of various maturities (technicals). And it’s not a particularly strong tool at that.

It’s the resulting interest rates that may or may not alter the economy, inflation, and the value of the dollar, etc. and not the quantities of reserves and Treasury securities per se.

And it is clear to me that the FOMC does not fully understand this.

If they did, they’d be in discussion with the Treasury about cutting issuance.

And, additionally, If they wanted the term structure of interest rates to be lower, they would simply target their desired term structure of rates by offering to buy unlimited amounts of Treasury securities at their desired rate targets, and not worry about the mix between reserves and Treasury securities that resulted. Which is what they did in the WWII era. And how they target the fed funds rate.

With today’s central banking and monetary policy with its own currency, it’s always about price (interest rates) and not quantities.

32 Responses

  1. Goes without saying – a very good description – but I said it.

    One observation:

    “First, recognize that, in fact, reserves, functionally, are nothing more than 1 day t bills.”

    I think there is one functional difference worth noting. Banks are free to sell 1 day or 30 day treasury bills to non-banks, altering their liquidity positions in doing so. Reserves on the other hand are “trapped” as banking system assets until the Fed “un-traps” them by draining. Or until currency demand reduces bank reserves, but that’s at the discretion of non-bank demand for currency, and happens ongoing regardless of QE injections.

    I think of this trapped characteristic also as encompassing “banks don’t lend reserves”. They also don’t sell reserves to non-banks. That’s a functional difference from Treasury bill holdings.

    That said, its more a quantitative functional difference as opposed to a pricing functional difference, so it doesn’t affect the pricing over quantity theme. But it’s a difference nevertheless.

    1. Could say:

      “First, recognize that, in fact, reserves, functionally, are nothing more than 1 day t bills issued by the Fed.”

      1. no

        reserves are issued by the Fed

        but they have the interest rate characteristics of treasury bills

        so they’re like t bills issued by the Fed, instead of by Treasury

      1. That’s going pretty wide, but OK.

        Clinton might say it depends on how the meaning of “functional” functions.


  2. I believe that QE is about both price and quantities. Total govt debt is about 13T, out of which Fed has about 3T. 3 years ago, Fed had about 1T. Fed balance sheet has now expanded to 3T. Your arguments are accurate only if the Fed balance sheet remained at about 1T. Fed has now effectively reduced the govt debt by about 3T.

      1. It matters. The total quantity of resources available to public is now reduced from 13T to 10T. It is conceptually similar to a corporation issuing 13T IPO and then subsequently buy back 3T thereby reducing the outstanding shares.

      2. I think my previous comment is incorrect. QE is similar to altering the capital structure rather than decreasing the capital. Before QE, it was similar to a private corporation with 12T loan and 1T outstanding shares. After QE, it is similar to 10 T loan and 3 T outstanding shares.

        Is my understanding correct ?

      3. sort of. correct that non govt financial assets are unchanged. just different maturities available for selection, with interest rates adjusting to keep them all indifference levels.

    1. At least this part can be agreed on:

      “the supply of loans is not the constraining factor, it’s the demand. Increasing the supply of loans won’t have much of an impact if firms aren’t interested in making new investments.”

      From there it is just a short step to admitting that the quantity of reserves, per say, is not a constraining factor on loan creation. As you and Mr. Mosler points out, it is the price, not the quantity. Actually, I think Professor Thoma might understand this, but is just incredibly sloppy with his language. After all, commercial banks must eventually obtain reserves for settlement purposes, no? Providing additional liquidity by increasing the amount of reserves in the banking system should lower these costs, correct? Perhaps Professor Thoma believes these lower costs will be passed on to borrowers. Or am I completely confused?

      1. “After all, commercial banks must eventually obtain reserves for settlement purposes, no?

        ***it’s an overdraft until they do, and an overdraft is a loan from the fed at a penalty rate.

        Providing additional liquidity by increasing the amount of reserves in the banking system should lower these costs, correct?”

        ***yes, in that the fed funds rate will trade up to approx the overdraft rate as the bank tries to obtain cheaper funding. to keep fed funds at its target rate the fed will offer cheaper funding via open market operations. it’s about price and not quantity.

        also, with today’s excess reserves this isn’t an issue like it used to be

      1. Interesting discussion and you explained that very clearly Warren, thanks.

        I’m curious though if its a problem that the yield curve hid 3 mon .14%, 6 mo 0.19%, 1 yr .26% but the (IOR) 1 day bills yield .25%. In consumer banking terms, its like a 6 month CD yielding less than than an interest-bearing checking account.

  3. Warren: Question for you, in soft currency economics you state that As long as the Fed has a mandate to maintain a target fed
    funds rate, the size of its purchases and sales of government
    debt are not discretionary. But what happens in the current situation when the fed funds rate is at zero, is the fed monetizing now?

    1. today with the ability to pay interest on reserves the fed uses that tool to support its fed funds target.

      so with that tool the fed can buy all the securities it wants without the ff rate falling below its target rate.

      and yes, with a 0 rate policy that’s true at well even when it can’t pay interest.

  4. While I understand the technical aspects and do agree that QE is just issuing 1 day T bills, I am still at odds with the idea that this doesn’t have any impact anywhere. Otherwise, why not have every Central Bank (with fiat currency power) just do QE’n’ all the time there is slack in the economy and get all this done and over with. Any time there is a recession, just QE it away. If it takes $8 trillion (and because of slack in labor and capital, there acmn’t be inflation), do it.

    I still see a couple of problems (and I am sure there are more):
    1. You are creating cash but not any “value”. Inefficient channels prevent it from going to the public (because there is no demand) and instead excess reserves are being speculated upon to drive up assets in a bubble. Creating cash without adding “value” (or it being used to add “value”) to the society will result in inflation.
    2. Channeling the cash efficiently is not practically possible (this is sort of a corollary to 1 above). End result is overvalued assets and bubbles.
    3. While technically QE just retires longer term Treasuries with newly minted cash and therefore doesn’t actually change govt liabilities, my question is doesn’t the newly minted cash retire old Treasuries which were issued to finance productive assets (yes, I know Treasuries don’t actually “finance” anything and yes I also know that the recent issu of Treasuries could have been deployed for similarly futile unproductive asset bubbles, but take this argue long back enough and you will get the idea), then again you are creating $ without it being deployed towards productive assets. Similar point as # 1 above. Where is the “value creation”?
    3. What about carry trade effects, excess reserves lead to $ flooding into emg mkts and commodities creating inflation there in turn causing price rises here as well (at least in certain commodities)?

    The single biggest issue I have with all this is still the same one that I have had since a year or so ago – if it is so simple, what is preventing it from happening? Why not print $5 trillion and avoid this whole damn recession? And the next step to this, if you can indeed solve the problems this way, what prevents citizens from not working, continue speculating and being bailed out by central bank printing every time there is a bubble i.e. why should I actually produce or do any work? I know that the central bank can just print and solve any temporary recession, so I just sit and speculate. I might lose temporarily while asset prices fall or crash, but secure in my knowledge that the central bank will just print, I have no reason to work. This then leads to sloth and eventual decay of the citizenry and the country – doesn’t it?

    What am I missing?

  5. One correction–we need to start talking about the size of the national debt that is at issue correctly. QE, interest on the debt, Treasuries, etc., have nothing to do with internally or govt owned debt such as the trust funds. Thus, $13T is NOT the correct number, even though it’s the one everyone uses.

    The numbers that matter as of 2010q2 (from St Louis Fed’s FRED) are thus the following:

    Treasuries owned by Fed=$777B
    Treasuries owned by private investors=$7857B

    It may not seem like a big deal, but note that the Tsy’s held by pvt investors is the only part that the Tsy actually services with payments to the non-govt sector. And that debt is less than 60% of GDP, not the 80% to 90% figure we often hear when $13T is used. I know, as an MMT’er that debt ratios don’t really matter, but (1) being precise matters when there is fear mongering everywhere around us and (2) because debt service IS what matters, it’s important to know which measure of the debt is relevant for that.

  6. There’s an interest article on Seeking Alpha today about QE2 which uses an MMT (and mentions Warren by name) framework to explain the Fed’s decisions. It makes a compelling case that QE2 is meant to buy up US debts in order for the interest payments to be refunded back to the Treasury and provide room for more tax cuts/spending increases. Of course this is not strictly necessary, but with our self imposed limits on spending it makes sense. Here it is:

  7. Excellent post, with clarity.

    How did QE 2009 differ from the the QE 2010? Am I correct, that QE 2009 was about buying a broader mix of securities (not only Treasuries, but agencies and MBS etc) from agencies (and not the private banks, as is the intention for QE2).
    QE 2009 involved POMOs of $1.25 t of Agency MBS, $300b of Treasury obligations and $200b of Agency Bonds.
    And QE 2010 will look more like POMOs of 3month Tbills from bank portfolios ===> increase in bank reserves.

    So, in a way different but not really? What say you?

Leave a Reply