Liquidity hoarding and the end of QE

Isn't this interesting?

(lines and QE annotations mine)

Every time QE is announced, yields rise: when it ends, they fall. And no, this doesn't just affect the 10-year yield. The same basic shape can be observed on just about any maturity over 1 year (short-term rates are propped up by the positive IOER policy).

I've written about this before, and concluded on that occasion that the rise in yields was due to the closed-end nature of previous rounds of QE bringing forward sales that would not otherwise have happened and encouraging carry capture strategies due to raised inflation expectations. I expected therefore that if QE was continued for long enough, or announced in a way that indicated no definite end, yields would fall as expected rather than rising. But it seems this is not the case. The current round of QE was announced in September 2012 with no end date for purchases. But yields started to rise soon after it was announced: admittedly they rose more gradually than in previous rounds, but rise they did. And they continued to rise until December 2013, when the Fed announced its taper. Since then, yields have been falling.

Now of course there may be all manner of explanations for this. Correlation doesn't indicate causation, ceteris is not necessarily paribus, and all that. Deutsche Bank came up with eleven reasons for falling US Treasury yields, none of which mentioned the taper, and David Ader of CRT Capital Group produced another eleven that didn't mention it either. And I am sure there are plenty more reasons that don't include tapering.

But I have a theory. I think it IS at least partly due to the taper. And the reason is the effect that QE has on global liquidity.

QE increases liquidity in the financial system, or rather in the regulated part of it*, by buying assets that carry some form of risk - usually duration risk and interest rate risk. This has the effect of moving risk from the financial system to the central bank's balance sheet. The consequence of this is that the financial system is absolutely awash with the safest form of safe asset, namely cash, and rather short of slightly more risky cash substitutes. Those who want safety have no reason to buy longer-dated treasuries when there is so much cash around, while those who want yield will still prefer riskier assets. Rather than depressing yields on longer-dated treasuries, therefore, QE actually raises them as investors substitute cash for treasuries**.

When QE stops, whether suddenly or gradually, there is of course no immediate withdrawal of liquidity. But the sudden removal of the INCREASE in liquidity gives the impression of a drought. It's like someone washing their hands under a running tap instead of in the sink: when that tap is suddenly turned off, or the flow through it is restricted, the washer thinks they have run out of water, even though there is an entire sink full because of the previous flow. This is what is happening in financial markets. The Fed is turning off the QE tap.

A year ago, when Bernanke first announced that the Fed intended to end QE, markets thought the tap would suddenly be turned off. And they panicked. So that's not how the Fed is doing it. They are turning off the tap very, very slowly, giving lots of time for market participants to find other ways of managing liquidity. And, prodded by regulators, market participants are doing so. They are building up substantial liquidity reserves in the form of safe liquid assets. This is happening across all types of market players: governments, especially EMs, are building up FX reserves; corporates are building up cash and near-cash holdings on their balance sheets: banks are vastly increasing their liquidity buffers; and asset managers are increasing their safe asset holdings to give them collateral so they can meet day-to-day liquidity needs, for example for cash margin calls on derivatives traded through central clearing houses.

To be sure, all of these players have been building up liquidity reserves for quite a while - in some cases, since well before the financial crisis. But the withdrawal of QE makes this need more acute.  As I explained here, the age of easy money is over, at least for the moment.

So market participants are increasing their purchases of safe assets in order to improve their liquidity. All classes of goverment bonds are involved: after all, if liquidity drought is the perceived risk, a future claim on Government cash that it is obliged to honour is valuable. As is anything that is guaranteed by government and readily tradeable for cash. It's worth remembering that for corporations, insured deposit accounts don't offer much in the way of safety because deposit insurance limits are too low: government bonds are nearly as liquid as an insured deposit account and benefit from an unlimited government guarantee. Corporations who wish to self-insure their own cash flows, therefore, will want to hold government bonds or equivalent safe assets as stores of liquidity. Banks, too, hold government bonds as liquidity buffers in addition to, or as substitutes for, reserves.

At the same time, we have capital rebounding back into Western safe assets from the troubled EM countries and a slowing China. And we have the beginnings of a slowdown in Germany, too. All of these combine to depress yields on USTs as fearful investors seek safety, fearful corporations and financial institutions hoard liquidity and the Fed continues its taper.

Government bonds are not the only safe assets whose price is being pushed up. Property, too, is rising in price. In Europe it is not the most liquid of assets, so the purchasers tend to be people looking for yield rather than liquidity hoarders. But in America, where the mortgage market is largely securitized, the price of MBS is soaring. Property itself may not be liquid, but its derivatives are.

If I am right that the cause of the current fall in safe asset yields is primarily liquidity hoarding due to the Fed taper, then eventually yields will reach equilibrium. But.....there is of course another explanation. Yields on safe assets have been falling for over thirty years:

Could it be that all QE3 did was temporarily prop up yields, and now that it is being be removed, yields are simply reverting to their previous trend?

I'm sure this is what Larry Summers would say. Welcome to secular stagnation.

Related reading:

Weird is Normal 
Rediscovering IS/LM - Pieria
The negative carry universe -FT Alphaville

* QE can actually create a liquidity drought for non-banks who rely on the same safe assets as collateral for repo financing as the central bank is buying. This is partly why the Bank of England used a collateral swap, rather than more QE, to provide banks with cheap funding for lending: increasing the volume of T-bills in circulation and encouraging banks to use them in the repo markets eased the safe asset shortage and therefore made liquidity conditions easier for non-banks as well as banks.

** Just to remind you: yield is the inverse of price. If yields rise, prices fall.


  1. Hi - Was just wondering if prehaps you were overthinking this. (Occams razor and all that?) In my mind there is no great mystery here, its simple economics : When central banks do QE, it pushes up GDP growth and inflation expectation, which push up expectations for future interest rate hikes, which cause the yields on the long end of the market to rise.

    Remember these are huge markets, QE has some effect on the flow, but in general price is determined by the stock, not the flow...


    1. I agree this explains the rise, but it does not explain the fall. The excess monetary base that is believed to cause inflation is not withdrawn when QE ends, so there is no reason for inflationary expectations to fall. And as the purpose of QE is to stimulate the economy to achieve long-run growth, it is irrational to assume that its effects only apply as long as QE continues.

      QE affects the stock as well as the flow.

    2. Frances, you say "excess monetary base that is believed to cause inflation". Is it possible that this "excess monetary base" has no direct relationship to the money supply which would be related to inflation?
      The money supply in the economy is created only when deposits increase in the banking system and hasn't the QE liquidity been created as excess reserves, not deposits?
      Those arguing that the excess reserves can be "released" into deposits are not understanding how banking works. Reserves have nothing to do with how much credit is created which is how the money supply expands. Credit is issued based on assets as security - if there are assets (free of "liens") and there is a need for more money then banks can lend against those assets without regard to central bank reserve status.
      When there are excess reserves then no additional reserves are created; when there are insufficient reserves the central bank simply creates more against the asset security backing the increased credit.
      This makes the argument that the excess reserves are an overhang, which can be released to fuel inflation, a defective concept. What is alarming is that so many strident voices are expressing that concept, including some central bank governors. No wonder there is so much confusion when some central bank governors don't understand the basic mechanics of money and banking.
      So, with the system as I have described it, when past credit excesses still remain in the system the asset base available for new credit is diminished (think underwater mortgages for one simple example). Not only is the demand for new credit diminished but there is an excess of issued credit above and beyond the value of the asset liens. That creates a deficiency in the central bank reserve account which needs to be "made up".
      This is one way to look at QE: making up reserve deficiencies. They show up as excess reserves simply because the financial system accounting is not accurately reflecting the bad debt on the books. Mark-to-market accounting (if accurately done) would create bank insolvency and that would consume all the excess reserves (and possibly more than currently exist).
      Rather than a monetary overhang it is possible that the excess reserves are currently insufficient to cover what I would call proper accounting by the banks.
      So in the scenario I have outlined there is little pressure for higher interest rates because there is little pressure for NET ADDITIONAL credit because credit still on the books is subject to "writedown" for still sometime in the future.
      The situation we are in is debt deflation. This was described by Irving Fisher in the 1930s - possibly one of the most important macroeconomic principles ever developed - and possibly the most overlooked and least understood factor in modern finance.
      With this situation there is little demand for new credit, the "market" rate of interest is very low and current rates may well be too high.
      You suggest in your article that interest rates might go lower (revert to previous trend) instead of higher as almost everyone is forecasting. You mention Larry Summers which implies a reference to negative interest rates. Yes, that is a logical extension of these arguments. But wouldn't a more straightforward solution involve honest accounting, writing down bad bank assets and removing the giant hidden suction pump that is demanding excess reserves to prevent systemic collapse?
      The problem with that, of course, is that "apparent wealth" (not "real wealth") is destroyed and a vast lucrative compensation scheme in finance is massively cut back. These interests are represented by powerful political agents who prevent the needed actions from happening.
      I apologize for an overly long rant in your comment thread.

    3. John,

      Great comment, and an excellent explanation of why removing QE would cause liquidity problems. I think Izzy Kaminska has written about this - I'm sure I remember something about "faux capital" that could not be distinguished from real capital. And I agree. We are in slow debt deflation, and I think it will probably last for a generation or more. Financial repression and capital destruction. The problem is that destroying the faux capital left over from the years of excessive credit creation inevitably also destroys (or prevents formation of) real capital.

  2. Frances, I'd like to confirm your thesis: There are many institutional investors who are seeking risk-free assets, and for them the QE induced plentiful cash is just a substitute for short-term government bonds (though some prefer government bonds since it provides explicit insurance that a bank cannot in large denominations). As QE winds down, the supply of cash will fall, and will push more safety seeking investors to purchase government bonds.

    If that is correct, then for yields to fall the demand for bonds by risk-free investors must exceed the former QE level of demand by the Fed. If the increase in bond availability equals the decrease in cash, then it is a wash and there is no long-term change in yields. I don't see US banks needing more liquidity, unlike some large European banks. On the other hand, tremors in China should cause many developing nations to increase their Treasury reserves.

    If the Fed raises Fed Funds (to say 100 bps in Q4 2015), then would that create a carry trade on the Euro, as a consequence of the japonification of Europe? If so, then that European demand would push Treasury yields lower.

    I keep warily looking at JGB yields of 1% and wondering why we aren't going there, which makes me the odd man out in my circle.

    1. Kent,

      Lots of interesting thoughts here. The FT today published an article saying that the bond rally is caused not only by increased demand - as I suggest - but also by falling supply as governments restrict bond issuance.

      I don't think the Fed has any intention of raising the Fed Funds rate. US fundamentals don't support this. I would guess that they are looking to replace QE rather than end it. We will see other forms of easing in due course, I think.

      I agree with you about JGB yields. The long-term trend for USTs is the same - indeed it is for all major govt bonds, really.

  3. There’s nothing unexpected about long rates rising at QE announcements and dropping at QE ends. Classical theory says money supply growth causes inflation and increases long rates. You could argue that all we are seeing is that theory, or belief in it, working.

    That said, I doubt it. In my opinion, bond markets have ignored stagflation warnings and dismissed the classical theory in favor of the liquidity trap theory, which says additional base money at the zero bound has minimal effect on inflation. I think Mikhail is right that the most important factor has been long-term growth expectations, but I'm not so sure those have turned on QE.

    I don’t think there’s any one single explanation for the multiple shifts in market sentiments on long-term growth over the past several years. And I don’t think they can be tied so neatly to QE starts and stops. Your attempts to do so are somewhat artificial. On one hand you date episode starts to official announcements rather than to actual starts of QE. On the other hand you date episode ends to actual ends of QE, which were fully anticipated and should have been priced in, or to the actual beginning of taper of QE3, which was very well telegraphed and anticipated. If I were looking for QE’s effects I’d look mainly at less formal announcements. For example the April 2013 Bernanke speech on taper was far more important than the well anticipated formal announcement of taper by the FOMC.

    Also, I think the announcement the launch of emergency liquidity operations in 2008 was far more important than the announcement of QE1, which wasn't new liquidity but only a shift of assets from emergency loans to Treasurys and agencies. The emergency liquidity was of course overwhelmed by negative sentiment and long rates plummeted anyway. The upturn of sentiment in early 2009 had nothing to do with QE, in my opinion.

    As for the direct impacts of QE on the liquidity of cash and Treasury markets, you need to apply the same logic to bond and cash liquidity. Just as the end of QE does not reduce the stock of cash, the start of QE does not reduce the stock of Treasurys held outside the Fed. The start of QE reduces the net issuance of bonds and increases the net issuance of cash. The end of QE reduces the net issuance of cash and increases the net issuance of bonds. When the market anticipates some rate of issuance of cash or bonds, less issuance than expected is felt as a tightening of liquidity. Central banks generally manage rates by slowing or accelerating the rate of net issuance of base money; actual contraction of base money supply is rare.

    If you want to examine how QE affects long Treasury liquidity, you would need to look specifically at Fed purchases of long Treasurys. Which was very light in QE1, light in QE2 and much heavier with Twist and QE3. You might find something interesting with such a study, but I think market sentiment on long-term growth has been by far the dominant factor.

    1. PS The relative importance of stocks versus flows depends on their relative scales. Back when excess reserves were minimal, changes in flows of base money issuance ruled the financial world. Since base money was already in massive surplus when QE1 was launched, that aspect of QE was never very important. Where QE and Twist mattered was as flows of asset purchases, as those were sometimes quite big relative to stocks, especially in mortgage bonds, and Twist and QE3's purchases of long Treasurys.

    2. Seems you did not notice that I talked about the announcement of start and end of QE. You should not judge my post by the chart alone.

    3. Thanks for the reply. I don't mean to condemn or be combative. You wrote: "Every time QE is announced, yields rise: when it ends, they fall." I'm pointing out those aren't logically consistent pivot points - you're saying markets buy the news before the facts when QE starts, but wait to buy the facts long after the news when QE ends. That doesn't persuade me that markets are turning on QE.

      If I understand your response correctly, you're referring to your discussion of the April 2013 Bernanke speech signaling taper, which as you say caused panic. I did notice that, and my point was that was the important turning point when the market understood QE would end. Yet rates didn't fall, as they should have by your opening thesis, they rose sharply.

      I don't understand why the December beginning of taper would drive anyone to suddenly change their minds about growth or inflation, or to hoard liquidity generally or sovereigns specifically. Besides being so well anticipated, this is more like slightly turning down a small faucet that pours into a big full swimming pool. I don't think long rates going down in a risk-on market can be called liquidity hoarding. Something else was afoot. Perhaps long rates were just settling down after an overdone taper tantrum. Sometimes unrelated events happen at the same time.

    4. Yes, it's the announcement that matters. That sentence is inconsistent.

      But I think the way it is done also matters. Bernanke's remarks caused panic, which the other "QE ends" announcements did not - including the real taper announcement in December.

      The taper tantrum is clearly an exception: yields rose sharply on Bernanke's announcement that the Fed was thinking about ending QE, whereas after previous "QE ends" announcements they fell, as they did when the Fed announced the ending of QE in December. I'm not entirely sure why Bernanke's remarks caused panic: maybe it was because QE had been billed as "QE infinity", so no-one was expecting it to end. Be that as it may, the Fed then DIDN'T end QE and rates went back to their previous gentle upwards path. During that time there was a lot of market chatter about tapering, so as you say the ending of QE was most likely priced in, which is why there was no panic when it actually happened.

      You've focused exclusively on one event and ignored the repeating pattern. What is your explanation for the evidence that QE is associated with rising rates, not falling rates, and after the ending of QE rates revert to their previous secular downward trend?

  4. PPS Oh, and QE also mattered to supplies of bank deposits. Banks will miss that.

  5. Both QE2 & QE3 both produced negative real-gDp figures (1st qtr of 2011 & 1st qtr of 2014). The payment of interest on excess reserves is responsible. Shifts in savings/investment accounts from the non-banks to the commercial banks exacerbated the trend (impounding monetary savings, decreasing the supply of loan-funds, depressing economic activity, & counteracting the Fed's open market operations of the buying type).

    I.e., the expansion of Reserve bank credit during quantitative easing was inadequate to offset the slow growth in non-bank lending/investing (pre-Great Recession, 82% of the lending market).
    The 24 month rate-of-change in proxy for money flows decelerated sharply after both QE2 & QE3 programs.

    Unless money (& money flows), expand at least at the rate prices are being pushed up, output can’t be sold and thus jobs will be cut (i.e., there is insufficient upward & downward price flexibility within our economy, e.g., “sticky wages” to counteract a contractionary money policy).

    The problem is that the Fed has emasculated its “open market power” (its ability to fine tune the money stock). Whereas between 1942 & 2008 the CBs minimized their non-earning assets by buying, e.g., T-bills, now the Fed accommodates the bankers with a remuneration rate that exceeds all money market rates (i.e., exceeds all returns in the wholesale funding market on the short-end segment of the yield curve).

    So POMOs between the Reserve and commercial banks only affect excess reserves (become just asset swaps). Before the CBs would buy short-term securities (increasing their liquidity reserves & the money stock). After the introduction of the payment of interest on reserves the CBs now simply hold higher yielding IBDDs.

    The Fed’s policy rate (credit control device), induces dis-intermediation among just the non-banks. It provides the CBs with the option to outbid the NBs for loan-funds (“specials”). This is exactly the same wholesale funding problem that existed with Reg Q ceilings: where during the 1966 S&L credit crisis, the CBs outbid the NBs for loan-funds).

    Unless CB lending/investing offsets the decline in NB lending/investing, then aggregate monetary purchasing power will be lost.

  6. Interest rates are determined by the supply of, and demand for, loan-funds - not the supply & demand for money (or Keynes' fallacious Liquidity Preference Curve). Economists simply don't know the difference between money creating depository institutions (commercial banks), & financial intermediaries (non-banks), between liquid assets (e.g., divisia aggregates), & money. Money is the measure of liquidity (bank debits). And "reserves are driven by bank debits" - Dr. Richard Anderson

    The non-banks are not in competition with the commercial banks. The CBs could continue to lend even if its customers ceased to save altogether. Money flowing through the non-banks never leaves the CB system. The NBs are the CB's customers. Thus all demand drafts originating from the NBs clear thru the CBs.

    Since Bankrupt U Bernanke destroyed NB lending/investing, the Fed has to offset the decline of these risk takers by expanding the CB system & bank credit at much higher growth rates.

  7. Your absolutely correct in pointing out the unexpected fall in rates after QE ends. If money was easy (i.e., roc's in AD was increasing), then the transactions velocity of money would increase following the "trading desk's" liquidity injections. The fact is that the payment of interest on excess reserve balances at its present level depresses economic activity.

    For an explanation of non-bank (non-inflationary), vs. commercial bank (inflationary), lending/investing (traditional vs. shadow) see:

    "Should Commercial Banks Accept Savings Deposits?” by Leland J. Pritchard, Edward E. Edwards, and Lester V. Chandler at the 1961 Conference on Savings and Residential Financing in Chicago, Illinois

  8. The best methods to eliminate any "noise", when comparing money flows (payments), with goods & services is to (1) use identical time periods, (2) to use the historical, distributed lag effect, & to (3) ignore the seasonally mal-adjusted data.

    Money flows are cumulative figures (i.e., temporary liquidity injections can later be "washed out", or put another way, price changes are transitory and "revert-to-mean" if liquidity is "sopped up" within the distributed lag period).

    Rates-of-change (roc’s), in short-term money flows (proxy for real-output), are always a mirror image of the seasonal economic inflection pattern (I.e., empirical evidence that roc’s in MVt = roc’s in real-output).

    First column = proxy for real-gDp. Second column = proxy for inflation:

    01/1/2014 ,,,,, 0.16 ,,,,, 0.34
    02/1/2014 ,,,,, 0.13 ,,,,, 0.38
    03/1/2014 ,,,,, 0.14 ,,,,, 0.32
    04/1/2014 ,,,,, 0.15 ,,,,, 0.33
    05/1/2014 ,,,,, 0.15 ,,,,, 0.39
    06/1/2014 ,,,,, 0.14 ,,,,, 0.35
    07/1/2014 ,,,,, 0.14 ,,,,, 0.30
    08/1/2014 ,,,,, 0.09 ,,,,, 0.26 projected short-fall in AD
    09/1/2014 ,,,,, 0.09 ,,,,, 0.27
    10/1/2014 ,,,,, 0.02 ,,,,, 0.23
    11/1/2014 ,,,,, 0.02 ,,,,, 0.22
    12/1/2014 ,,,,, 0.03 ,,,,, 0.16

    The deceleration in the proxy for real-output (from July to August) of 5 percentage points is somewhat ameliorated due to the seasonal downswing, but a fall from July to Oct of 12 percentage points is a recipe for a major sell off.

    Thus we already know that QE 3 failed. The acceleration in growth during the initial phase was not due to the Fed's "open market power". It was the reversal of unlimited FDIC insurance coverage (which incentivized saver/holders to move money back through the non-banks thereby increasing Vt).

  9. The money is going out of stocks and into bonds. Here's the reasoning: stocks already went up a lot in value and now even though they're still climbing the stairs, volume is down. This indicates dissatisfaction with these positions.

    The money now is looking for a place with better prospects and it may have found it in bonds. The secular trend is for rates to go to negative zero so the long bond offers both yield and asset appreciation. But once rates get to zero, it will be too late since yield will be non-existent and room for appreciation will be tight. Clearly, for those willing to board this train, the time is now.

    Basically, this train is parked on slight decline. As more and more pour into the bond train, frictional forces are overcome, wheels begin to roll, and the bond train descends. It takes interest rates along for the ride.

    Basically, the bond bubble is like any other. The bond promises yield and asset appreciation. This promise satisfies demand. The demand makes the promise self-fulfilling.


Post a Comment

Popular posts from this blog

WASPI Campaign's legal action is morally wrong


What really happened to Signature Bank NY?