Don’t get your hopes up

I haven’t had a chance to fully digest yesterday’s Fed action.  However since I have received many comments from people who think I should be enthused, I feel obligated to throw a little cold water on the celebration.

Yesterday the Fed announced that they would engage in quantitative easing, which is one of the three major steps that I had hoped they would take.  Unfortunately, it is the least important of the three.  Just to review, I suggested that a multipronged approach was essential:

1.  An interest penalty on excess reserves.

2.  A clearly spelled out target path for NGDP or the price level.

3.  Quantitative easing, if necessary.

4.  Keep expanding until market expectations signal NGDP growth is likely to hit the Fed’s target.

If the Fed had done both one and two, I don’t think the QE would even be necessary, without those other two steps, my fear is that we’ll just get more of the same—banks hoarding lots of interest-bearing reserves, which are now essentially T-bills in all but name.  So this “monetary policy” is essentially swapping one type of government debt for another.  (I’m sounding like Paul Krugman here.)

On the other hand the S&P did jump 2% on the news, and since it was partially anticipated (I wish I knew the odds) we can infer that the actual impact was substantially larger.  But the level of the S&P suggests that we are far from where we should be, and an indicator I put even more weight on, inflation expectations in the TIPS market, barely budged at 5 and 10 year maturities.  The five year figure is a mere 0.69%.   I’m not an expert on forecasting, so these observations are very tentative, but I certainly didn’t see the sort of strong market reaction one would expect after a credible policy announcement.

I also don’t know what to make of the numbers—$300 billion in Treasury notes, and even more in purchases of other types of debt.  Does this mean the monetary base will go up by this amount?  If so, when?  And if QE is such a good idea, why has the Fed reduced the monetary base by $250 billion over the past 10 weeks?  Will the $300 billion merely reverse that decrease?  My hunch is that all these questions miss the point.  Hamilton has a post where he says:

The Fed has declared pretty loud and clear that it is not going to allow deflation.  So here’s my personal investment advice: don’t bet against the Fed.

My first reaction was that Hamilton was wrong, that the TIPS market had bet against the Fed.  And recall that I view market expectations (however imperfect) as the best we have.   The more I thought about it, however, the more I concluded that Hamilton is exactly right, the market read this as an indication that the Fed will do whatever is necessary to prevent outright deflation.  Instead we are likely to get continued below target inflation, and NGDP growth that is far below target.  For that the stock market was thankful.  But having spent so much time studying market reactions to truly effective stimulus in the Great Depression, I was underwhelmed by yesterday’s response.  A policy that was expected to boost NGDP growth up to 5% would have had an explosive impact on equity prices, and might well have driven long bond yields up, not down.

I hope this post doesn’t sound too negative.  The Fed is moving (slowly) in the right direction.  The market reaction was positive.  However, it is also important that we keep pressing the Fed to undertake a truly credible policy initiative, which would also include limits on bank hoarding of reserves, an explicit target path for one of the nominal aggregates, and a promise to target the forecast (whether internal or market-based.)

By the way, in Mankiw’s blog today I noticed that he described how a “clever grad student” had proposed a negative interest rate on base money.  I wonder if this clever student reads my blog.  Or the two short papers I published this year in The Economists’ Voice.  To be fair, he also proposed that the plan be applied to cash, but the method was so convoluted (invalidating currency with certain serial numbers, chosen randomly, at certain dates) as to be totally impractical.  And of course Gesell proposed a similar idea back in the Depression.  But the plan is very easy to implement for reserves, and almost all of the base hoarding has been in reserves.  Why hasn’t the Fed done this already?

The idea in the Mankiw column is not politically or practically feasible, as it would turn Federal Reserve Notes into a giant Keno game.  But there are more pragmatic solutions to the worldwide crisis that might actually work, like charging interest on excess reserves.

P.S.  Dilip just sent me a post by Yves Smith, who knows more than I do about the details of the Fed plan.  I would just add one point; some of the discussion focuses on whether long bond yields might rise.  If you told me today that long bonds would be yielding 5 or 6% a year from now, I’d be thrilled, just as I’d be thrilled to hear that oil would be back up to $100.



45 Responses to “Don’t get your hopes up”

  1. Gravatar of Dilip Dilip
    19. March 2009 at 10:43

    I think Yves Smith is a ‘her’ 🙂

  2. Gravatar of ssumner ssumner
    19. March 2009 at 10:48

    Thanks Dilip, I changed the wording so it will work either way.

  3. Gravatar of JKH JKH
    19. March 2009 at 11:02

    “And if QE is such a good idea, why has the Fed reduced the monetary base by $250 billion over the past 10 weeks?”

    The reduction in the monetary base corresponds pretty closely with the reduction in foreign currency swaps (central bank liquidity swaps) on the asset side of the balance sheet:

    I wouldn’t necessarily view pulling back on this particular aspect of Fed asset expansion as inconsistent with a more general trend to quantitative easing. Foreign currency swaps are targeted uniquely at global dollar liquidity rather than domestic liquidity. The fact that European and other regional dollar liquidity has improved on a relative basis doesn’t necessarily mean the Fed shouldn’t continue with trend on its domestic course of action. In fact, the Fed must be relieved that the foreign swap facility has “worked” to the degree that foreign central banks are now paying down some of their utilization. This creates some comfort in the sense of some “unutilized capacity” in terms of positioning for yesterday’s domestic announcement.

  4. Gravatar of JKH JKH
    19. March 2009 at 11:10

    I’m curious.

    Why do economists seem to focus only/mostly on the monetary base implications of quantitative easing and the behaviour of banks as “principals” in “recycling the base?

    (I’m assuming the discretionary base effect shows up as increased bank clearing balances; not as increased notes in circulation)

    Why don’t you talk about the immediate broader money effect (e.g. M1 expansion) as well, where banks are “agents” for additional monetary activity?

    And why don’t you compare the two effects?

    (Maybe you do and I haven’t noticed.)

  5. Gravatar of Scott Sumner Scott Sumner
    19. March 2009 at 11:17

    JKH, I don’t see why the foreign currency swap matters. Either these swaps boost the base or they don’t. If the Fed wants to do swaps without boosting the base they can. But if you are right that swaps explain the drop in the base, why is this an excuse? Either changes in the base impact the stance of monetary policy, or they don’t. If the $250 billion drop in the base didn’t matter, then why would an increase matter?
    But monetary theory, and particularly the theory behind QE, says that increases in the base are what matter, not the particular assets that were bought to buy the base.

    You may well be right, as I don’t know as much about the balance sheet as you do. But I am skeptical of all these detailed balance sheet explanations. James Hamilton, who knows much more than I do, is very skeptical about whether the Fed’s balance sheet tactics have had any effect, he calls for straightforward QE, and says don’t worry about trying to influence spreads, etc.

    Regarding your other comment. I don’t think any aggregate is particularly reliable, including the base. I focus on market expectations. But the base is something the Fed can directly control. Some claimed the Bank of Japan was unable to increase M2, because of reserve hoarding. But no one can argue that about the base. So at least the base has the virtue of showing that the Fed is “doing something.” It is not doing enough, but that is another matter.

  6. Gravatar of JKH JKH
    19. March 2009 at 11:43

    Swaps certainly change the base, other things equal. A number of these Fed programs involve draw downs of credit facilities, where the draw down or repayment timing pattern is not necessarily at the discretion of the Fed in the short run, once the facility has been extended. The foreign currency swaps are one such facility. It is the foreign central banks that are choosing the timing of their repayments. I don’t think the Fed would simply view the repayment of a swap tranche as an urgent signal to replace the monetary base effect, at least up until now. It seems to me the quantitative easing effect of the various programs of the Fed up until now has been the consequence but not necessarily the objective of its qualitative or credit easing programs on the asset side, including the swaps. Even with the program announced yesterday, the parameters are set in terms of asset objectives, not net monetary base expansion objectives. I’m not even sure that could be described as quantitative easing in the sense that there is no announced objective for the monetary base (maybe there is and I missed it).

    My other point was simply that the first round effect of quantitative easing includes broad money supply expansion by the same amount as reserve expansion. Even if banks sit on their reserves, with no further effect, the public has additional M balances to circulate apart from that.

  7. Gravatar of Quantitative Easing Reaction Roundup « Random Musings of a Deranged Mind Quantitative Easing Reaction Roundup « Random Musings of a Deranged Mind
    19. March 2009 at 18:12

    […] Scott Sumner was somewhat less impressed. […]

  8. Gravatar of Jon Jon
    19. March 2009 at 19:31

    JKH is partly correct. As the swaps have been unwound the Fed’s balance sheet has contracted the monetary base has contracted. This accounts for the 250B roughly. But there has been an offsetting transaction.

    The treasury’s balances at the Fed (which sterilize the MB) have declined from close to 600B to roughly 300B now. Thus 300B of the MB that seemed meaningful was not; it had been sterilized.

    See (Red line is the declining sterilization)

    However, I think that JKH is also right to suggest that the currency swaps were never (domestically) stimulative because the depository institutions that received those funds (ultimately) do not engage in lending in the US market.

    All these points are made clear in the following:

    The monetary base is falling, but monetary less reserves is stable. The portion of the Fed’s balance sheet allocated to domestic channels is rising.

  9. Gravatar of JKH JKH
    20. March 2009 at 04:50


    You’re right. The unwinding of the swaps is directionally consistent with the unwinding of treasury balances. To be honest, I haven’t been looking at all the interrelated numbers closely, and forgot about that aspect. Treasury balances are a form of effective sterilization for what otherwise might be the same marginal amount of quantitative easing.

    The unwinding of treasury balances was announced as an intention roughly around the same time that payment of interest on reserves began. I view the payment of interest on reserves as a structural change that will facilitate exit from quantitative easing, rather than a material impediment to the effect of quantitative easing now. Not sure there’s a huge difference between paying 25 basis points and paying zero at this particular point. But the change communicates the fact that an effective floor will be in place for fed funds, when the fed funds rate eventually starts to move higher, even in the presence of any residual quantitative ease (i.e. excess reserves).

  10. Gravatar of David Pearson David Pearson
    20. March 2009 at 05:43


    Given your desired oil price level ($100), you might find the views of a market participant relevant.

    I am essentially betting that the Fed will progressively engage in balance sheet expansion. I am also betting that the reversal of this expansion will be extremely problematic. One might say I am “hedging” against this risk. The way I hedge is by buying gold and other commodities. As I and others engage in this behavior, commodity prices should rise, and so should inflation. I hope agree that this is the dynamic which would make the oil price rise far in advance of actual CPI inflation.

    Of course I’m describing the change in inflation expectations amongst market participants; a change which you want the Fed to engender. The problem is that hedging behavior is anticipative; perhaps highly so. Commodity prices could well rise before real GDP growth appears. The effect could well be a 3-5% inflation rate with little or no real growth. What does the Fed do in that scenario? Presumably you would have them nurture that inflation rate for a few years before scaling back stimulus. The problem is that market participants, seeing the Fed’s LACK of reaction, would further hedge against hyperinflation, causing inflation to accelerate without an impact on real growth.

    In the above scenario, the Fed would face a choice: force unemployment up, or risk expectations getting out of control. No “smooth” landing for stimulus, in other words. And yes, the Fed will attempt to control expectations by promising a long term inflation target. But that target will entail forcing up unemployment at a time when the long-term unemployed will be a significant societal problem.

    So in a sense I am the embodiment of “your worse enemy”. I won’t believe the Fed until it demonstrates a tolerance for high unemployment. As long as I don’t believe the Fed and stimulus remains, I will hedge. As long as I hedge, I both make the Fed’s job easier (at producing inflation) and vastly more difficult (at eliminating stimulus). So ironically your wish (of $100 oil) is my desire, but I don’t think you will be happy with the outcome.

    By the way, you might find it relevant that I successfully bet against the Fed in the other direction — that they would act too slowly to avoid nominal GDP declines. Having successfully exploited their weakness (overconfidence in the economy) on the downside, its natural that I don’t share your sense of confidence in their management of the inflation upside. The question is, given that you also think they committed a major blunder, why you would trust them with QE?

  11. Gravatar of Jon Jon
    20. March 2009 at 06:07

    “I view the payment of interest on reserves as a structural change that will facilitate exit from quantitative easing, rather than a material impediment to the effect of quantitative easing now.”

    I really don’t see how it will facilitate an exit. In particular, the policy is highly pro-cyclically. As the FF rates fall, the spread declines which encourages reserve hoarding. As the FF rates rise, the policy becomes extremely expensive and its natural to assume the spread will grow which will in turn be encouraging the banks to draw down their reserves–precisely when you mean to be tightening policy.

    Please elaborate on your thinking.

  12. Gravatar of Bill Woolsey Bill Woolsey
    20. March 2009 at 06:27

    There is nothing wrong with paying interest on reserves–as long as there is no notion that the interest rate should be fixed. My view is that the interest rate should be negative now, though there isn’t much point in making it so negative that banks will start to accumulate vault cash.

    On the other hand, they can raise the interest rate on reserves when they need to reverse course.

    I am slightly troubled that the Fed explained its purchase of agency mortgage backed securities in the context of supporting the housing market. This continues to suggest interest in influencing the direction of credit, in particular support of securitization–rebuilding the house of cards that collapsed.

    The purchases of long term treasures are appropriate.

    On the other hand, the notion that they will maintain their target for the federal funds rate (instead of dropping it to zero,) still leaves open the possibility that this “quantitative easing” will be steralized. That is, they will sell even more T-bills to fund their purchases on long term securities.

    Frankly, I would prefer them to announce targets for base money and various other measures of the money supply.

    And, of course, most importantly, to committ to getting nominal income back up to its previous growth path.

  13. Gravatar of JKH JKH
    20. March 2009 at 06:48


    The Fed tightens monetary policy (in general and normally) by increasing the Fed funds rate. In normal circumstances, the Fed faces no operational difficulty in enforcing a higher target funds rate, because it puts an effective floor in place by restricting the availability of excess reserves. Banks are forced to compete for available reserves in order to meet their reserve requirements. The Fed fine tunes aggregate reserve availability to steer the effective fed funds rate accordingly.

    With quantitative easing in place, the Fed faces an operational problem in enforcing a higher target funds rate if it doesn’t pay interest on reserves. Quantitative easing means that excess reserves are not restricted as normally is the case. If the Fed moves the target fed funds rate higher, and if banks earn no interest on excess reserves, and if excess reserves remain due to the residue of quantitative easing, banks will attempt to lend out excess reserves and in doing so force money market rates below the target funds rate. Therefore the Fed must pay interest on excess reserves in order to put a floor on the effective fed funds rate.

    The Fed may want to begin moving the target funds rate higher before it unwinds all of its quantitative easing. The judgement may be that monetary policy in general requires tightening even though specific sectors assisted by quantitative easing shouldn’t be abruptly unwound according to the same timing.

    So paying interest on reserves allows the Fed to move the target funds rate higher without being constrained by a requirement remove all residual quantitative ease from the system. The Fed may judge that the interest rate component of monetary policy needs to be tightened before the financial intermediation easing component is fully unwound. In other words, it gives them flexibility. BTW, Bernanke has referred to this desired flexibility resulting from the payment of interest on reserves several times in speeches and in congressional testimony.

  14. Gravatar of Alex Alex
    20. March 2009 at 07:46

    Have you read the latest from Krugman?

    “I’m not complaining; I think quantitative easing (it’s really qualitative easing, but I give up on trying to fix the terminology) is the right way to go. But we should go into it with our eyes open.”

    Now he is all in favor of QE.

  15. Gravatar of Bill Woolsey Bill Woolsey
    20. March 2009 at 07:58


    I believe you are fundamentally mistaken.

    Who cares what the Federal Funds rate does if it isn’t about the supply and demand for base money?

    Even if there were no reserve requirements, the Fed could manipulate the federal funds rate.

    Paying interest on reserves matters, not because it helps the Fed hit a target for the federal funds rate, but rather because it inceases the demand for base money.

    If the Fed wants to allocate credit, then raising the demand for reserves and then supplying credit to where they want it to go with acheive this.

    The point of “quantitative easing” is to create an excess supply of base money when the traditional approach/signal is no longer applicable because it has hit zero–the Fed funds rate.

    And, of course, it can actually be less than zero, but a bit.

  16. Gravatar of ssumner ssumner
    20. March 2009 at 08:30

    JKH#1, I see each part of your argument, but I can’t put it together in a way that makes any sense. You say that previously the drop in the base was a consequence of programs aimed at affecting balance sheets and credit markets–qualitative changes. And now the focus will be on quantitative easing (QE.) But the Fed doesn’t just create new theories out of thin air, if they now think that QE can affect AD, then they also thought so months ago (indeed Bernanke has made that argument numerous times.) So either they always thought the size of the base mattered (but were not targeting it), and now decided to target it, or they never thought it mattered, and still don’t. In the former case, they should never have let the base drop $250 billion, even if it wasn’t an explicit target. It doesn’t require any effort to prevent the base from dropping. On the other hand if they didn’t think the drop in the base mattered then, they probably don’t now, which raises the question of why attempt QE?

    Jon, I understand how a balance sheet move could be sterilized (by preventing it from impacting the base), but what does it mean for the base itself to be sterilized? The only thing I can think of would be if the Fed pays interest on reserves (as Bill mentioned), but that doesn’t seem to be what you are referring to. Aren’t bank reserves a completely homogeneous asset–each dollar of reserves is like any other? In that case how can part of reserves be sterilized, except by interest payments (or obviously reserve requirements?) I don’t see why it makes any difference which asset was acquired to inject the new base money, at least in terms of the sterilization issue (it may matter in other ways.)

    JKH#2, The Treasury balances are not counted as part of the base, are they?

    The reason a quarter point interest on reserves matters, is that T-bills yield 0.19%. Reserves are both more liquid and higher yielding than T-bills. No wonder so much is being hoarded. Now change the rate to negative 2%, or negative 4%.

    David, I should explain my $100 oil number. I don’t think they should target oil nor do I favor enough inflation to get $100 oil. We are not going to get 100% inflation, even if the indexed bond market is a bit off. I meant that if the world economy recovers strongly, then REAL oil prices would rise sharply. My point is that all the markets are expecting low inflation, and for good reason. They pay much more attention to the Japanese experience than the German hyperinflation, precisely because it is much more relevant (although as I’ve said, I think Bernanke is committed to doing a bit more than the Japanese.) I agree that there are some supply side problems, but I still think 5% nominal growth would mean more real growth than most people seem to think, even including some of my strongest supporters, such as Tyler Cowen. If Obama’s policies lower our long run growth to 2%, that would be a mind-boggling failure–and would still only mean 3% inflation. And as for short run, well that’s where you use the standard AS/AD model with monetary shocks, which affects output more than prices in the short run (especially in a deeply depressed economy.) The supply problems are roughly where we were in mid-2008, if you want to visual the supply side problems from the subprime crisis, in other words a roughly one percent rise in unemployment (which is obviously now headed for 10%.)

    Jon#2, I agree, it seems procylical.

    Bill, I agree with all your points, with one slight exception. I don’t think they need explicit targets for the base. If they gave us a NGDP target, and negative interest on excess reserves, I think it would get the job done. Everyone assumed I was gung ho for QE, and I suppose I talked a lot about it as a sort reductio ad absurdum argument that there was SOME monetary policy that would always be inflationary. But I would be more comfortable with much lower reserve holdings via an interest penalty, and then a smaller amount of QE could get the job done. BTW, the Fed will be buying these 10 year bonds at such a high price that they may well lose money in future sales, but I still favor QE if that’s all they are offering, because it’s better than nothing.

    JKH#3, Here you explain things very well. So let’s suppose the micro problem of bank liquidity requires massive reserve holdings (this is for the sake of argument, I actually doubt it) And also assume the macro problem of insufficient AD has been solved, as AD is expected to grow at around 4-5%. Then yes, there is an argument for positive interest on reserves. The mystery is why did they start this program in early October, precisely when the macro problem not only wasn’t solved, but was getting dramtically worse. Yes, had they not paid interest in early October, the fed funds rate would have fallen below the 2% target, down to zero. The puzzle is why the Fed would have thought that was bad. Weren’t they far below their target for expected NGDP growth? If not, why was the Fed calling for fiscal stimulus?

    Alex, Thanks, I’ll check it out.

  17. Gravatar of Alex Alex
    20. March 2009 at 09:08


    You are well come. But in the end I don’t believe that QE is what is going to get money moving around. What will get money flowing again is the huge increase in the nominal liabilities of the US government (M+B). Here is a graph I did today, the sources are FRED and the US treasury website.

    US Liabilities have gone up by $2T (or 33%) y/y in February 2009 and more increases are expected to come soon. My first guess would be that this would generate a 33% increase in the price level but since the Treasury/Fed are using part of the money to buy private sector assets then the net change in liabilities is somewhat lower. The final effect on the price level will depend on how much they spend in this assets and how they end up performing.

  18. Gravatar of JKH JKH
    20. March 2009 at 09:21

    Bill Woolsey,

    Not sure I understand all your points, but it looks likely we’re in fundamental disagreement.

    The Fed’s monetary policy in normal times is all about the fed funds rate, not about base money. Base money is derivative issue. In particular, it’s a subset of base money – bank clearing balances at the Fed – and a subset of that subset – excess reserves in the form of excess clearing balances – that the Fed controls in order to control the path of the effective fed funds rate relative to target.

    We are now in extraordinary times, not normal times. The Fed is using its balance sheet for extraordinary financial intermediation. There’s a lot of rubbish around in my view relative to so called quantitative easing and the difference between quantitative easing and the other type of easing, which Bernanke refers to as credit easing and Buiter refers to as qualitative easing. But quantitative easing involves an increase in the monetary base and qualitative easing involves moving to higher risk assets. I know there are precise definitions that some insist on for these terms but I think that’s part of the problem in that it obscures the real issues. As far as I’m concerned, the Fed’s been involved in credit easing since the first of its extraordinary credit programs, and in quantitative easing since bank clearing balances started to increase last fall. Wednesday’s announcement is simply another tranche of programs that involve both credit (mortgages) and quantitative easing (as a by product).

    To be honest, I haven’t been looking closely at the recent Fed announcements. But I’ll bet none of them say that the OBJECTIVE is to increase the monetary base. An increase in the monetary base is a by product of the actual objective, which is specified as an asset purpose for each program. The monetary base, specifically bank clearing balances in this case, is merely a necessary means to an end of asset expansion and an increase in the Fed’s role in financial intermediation.

    Banks don’t need clearing balances to lend. The banking system creates deposits from credit. In normal times, individual banks that are short clearing balances relative to any new credit they want to create will rely on liability management. The system as a whole doesn’t have to rely on anything, including the supply of clearing balances. The system creates its own funding from lending. In normal times, the purpose of clearing balances as far as monetary policy is concerned is to provide the Fed with a mechanism for the control of the effective fed funds rate. Some of this is more evident when you consider that Canada has a zero reserve requirement.

  19. Gravatar of Jon Jon
    20. March 2009 at 09:33

    Although you addressed this question to JKH: “The Treasury balances are not counted as part of the base, are they?”

    No, they are not. As you can see my reconstruction of the monetary base (shown in blue) closely follows the official monetary base and I explicitly subtract the treasury balances.

    “Jon, I understand how a balance sheet move could be sterilized (by preventing it from impacting the base), but what does it mean for the base itself to be sterilized?…”

    Lets be clear about something: the monetary base is a derived timeseries; it does not appear explicitly on the Fed’s balance sheet. Consequently, in the parlance of the St. Louis Fed, the treasury balances are not included in the monetary base because banks cannot opt to use them as if they were excess reserves. The treasury balances are instead in the custody of the Fed directly as if they were reserve balances.

    So I am at fault for the confusion, as I do not think of the monetary base literally as the St. Louis fed does. i.e., I view the monetary base as being closer to the actual monetization (i.e., the headline Federal Reserve Credit number). Historically, this matches the monetary base almost exactly. Thus I speak of sterilizing the base–i.e., holding money at the Fed other than that noted in the reserve balances. My justification for this is that the Treasury deposits are substantially like reserves–they are a wound spring–that the Treasury as an independent agent _could_ choose to release.

    As regards the swaps and whether monetary base is homogenous. I think the answer is clearly no. I.e., dollar currency in circulation in Africa has very little do with MV=Py from an American perspective. More so when you count M as the global dollar circulation but look at Py as GDP. Economists are wont to gloss over this because a given currency tends to have a common market, but that’s a horrible model for the dollar.

  20. Gravatar of JKH JKH
    20. March 2009 at 09:49


    Some of my response to Bill Woolsey responds to some of your points. I view the expansion of bank clearing balances (“quantitative easing”) as the secondary effect of what the Fed is trying to achieve. The Fed is trying to get credit flowing with its various programs. In all cases, including Wednesday, it has announced asset targets, not monetary base targets, as far as I know.

    In the case of the $ 300 billion treasury program announced Wednesday, this is the only program which arguably qualifies as pure quantitative easing under conventional definition. It qualifies, not because it is expanding the monetary base per se, but because it is an asset program that isn’t moving higher up the risk curve than normal for the Fed, with a secondary base effect. And the purpose of the program is to influence market rates and thereby influence mortgage rates – not to increase the monetary base. Again, the monetary base effect is secondary.

    So, the fact that the Fed is willing to pay interest on reserves is in a way consistent with the view that the monetary base effect of all these programs, QuantE or CreditE, is secondary.

    QuantE is a facilitator for CreditE – when CreditE involves balance sheet expansion as well as balance sheet recomposition. Again, the only exception to this so far is the $ 300 billion treasury program, but even there, the true objective is to influence the mortgage market, which is ultimately CreditE type of effect.

    Treasury balances are not part of the base.

    I agree 25 basis points makes a difference to the incentive for treasury bill purchases. But I don’t think this is the Fed’s primary concern. Banks could always move out the treasury curve to get a pickup beyond 25 basis points. Also, the payment of interest on reserves as a feature of quantitative easing, in addition to putting a floor on the Fed funds rate (especially for future consideration), puts in a lower bound of sorts on the potentially adverse effect of excess reserves on bank interest margins. The Fed has a vested interest in supporting banking system interest margins at this stage. It doesn’t want to be unnecessarily punitive. There is such a colossal amount of excess reserves sloshing around the banking system that paying interest on a systematic basis relieves banks of the operational pressure of going out into the market to make themselves whole simply for 25 basis points in this case.

    “Yes, had they not paid interest in early October, the fed funds rate would have fallen below the 2% target, down to zero. The puzzle is why the Fed would have thought that was bad.”

    I don’t think that’s quite the puzzle. If there is a related puzzle, its why didn’t they drop the target funds rate back then to where it is now. That’s how I would argue it if I did argue it (which I’m not necessarily), consistent with my view on the 25 basis points above.

  21. Gravatar of JKH JKH
    20. March 2009 at 10:51

    I took the time to do a little digging.

    Here is Bernanke’s key statement on the effect of the payment of interest on reserves:

    “Importantly, the management of the Federal Reserve’s balance sheet and the conduct of monetary policy in the future will be made easier by the recent congressional action to give the Fed the authority to pay interest on bank reserves. Because banks should be unwilling to lend reserves at a rate lower than they can receive from the Fed, the interest rate the Fed pays on bank reserves should help to set a floor on the overnight interest rate. Moreover, other tools are available or can be developed to improve control of the federal funds rate during the exit stage. For example, the Treasury could resume its recent practice of issuing supplementary financing bills and placing the funds with the Federal Reserve; the issuance of these bills effectively drains reserves from the banking system, thereby improving monetary control. As we consider new programs or the expansion of old ones, the Federal Reserve will carefully weigh the implications for the exit strategy. And we will take all necessary actions to ensure that the unwinding of our programs is accomplished smoothly and in a timely way, consistent with meeting our obligation to foster maximum employment and price stability.”

    Sound familiar?

    The above is excerpted from his recent speech on February 18, 2009, “Federal Reserve Policies to Ease Credit and Their Implications for the Fed’s Balance Sheet”.

    From the same speech:

    “The various credit-related policies I have described today all act to increase the size of both the asset and liability sides of the Federal Reserve’s balance sheet. For example, the purchase of $1 billion of GSE securities, paid for by crediting the deposit account of the seller’s bank at the Federal Reserve, increases the Fed’s balance sheet by $1 billion, with the acquired securities appearing as an asset, and the seller’s bank’s deposit at the Fed being the offsetting liability. The quantitative impact of our credit actions on the balance sheet has been large; its size has nearly doubled over the past year, to just under $2 trillion.”

    Note the logic of cause and effect – from credit expansion to liability (i.e. monetary base) expansion. Nowhere in this speech does Bernanke use the term “quantitative easing”. The reason he doesn’t is that it’s neither the policy nor the objective.

    The speech is here:

    And there is no fundamental difference in the way in which Wednesday’s presumed QE announcement was couched:

    “To provide greater support to mortgage lending and housing markets, the Committee decided today to increase the size of the Federal Reserve’s balance sheet further by purchasing up to an additional $750 billion of agency mortgage-backed securities, bringing its total purchases of these securities to up to $1.25 trillion this year, and to increase its purchases of agency debt this year by up to $100 billion to a total of up to $200 billion. Moreover, to help improve conditions in private credit markets, the Committee decided to purchase up to $300 billion of longer-term Treasury securities over the next six months.”

    Again, these are asset objectives – absolutely nothing about quantitative easing as a policy or a policy objective. There is no difference in the policy mode here other than new asset categories for balance sheet expansion. And note the use of the word “further”.

    Quantitative easing is a policy description about central banks that the Fed at least has never used and never announced.

  22. Gravatar of ssumner ssumner
    20. March 2009 at 11:38

    Alex, The reason I don’t buy the total federal liabilities argument is that it doesn’t seem to fit the facts. As Zimbabwe has found out, money financed deficits are far more inflationary than debt financed deficits. And closer to home, in the early 1980s we did an experiment of really tight money and really expansionary fiscal policy. If total liabilities mattered inflation should have risen sharply. It fell sharply.

    Jon, Your reply clarifies things a bit. The currency in Africa comment is off-topic, as I had argued that reserves were homogeneous, not cash. But maybe I worded the argument incorrectly. My thought was that there aren’t different types of reserves. You wouldn’t say Bank One has “type A reserves” injected through swaps, whereas Bank Two has “type B reserves” injected through bond purchases. So in that sense my argument would apply to currency. I understand that where the money is located has implication for velocity, but not how the currency was injected.

    JKH, There is a lot to respond too, and unfortunately I think you are on the wrong track. Before criticizing your views however, let me say that I do believe you have accurately characterized Bernanke’s views, and you and Bernanke both know a lot about this, but I still think the “credit approach” is wrong.

    1. In my short course on money post I argued that the fed funds rate is always overrated, and has little to do with how monetary policy is transmitted. A change in the base causes a change in the fed funds rate, not the other way around (otherwise the fed funds market could not reach equilibrium.) The fed funds rate is only for short term loans. Highly expansionary policy that massively increases the MB and leads to high inflation may not reduce the fed funds rate. Monetary policy works by changing the future expected path of the base, which changes the future expected path of nominal GDP. The fed funds rate changes as a by-product, and is highly visible, so people fixate on it. But it has little importance. The central banks of the world think it is important, but they delude themselves. When inflation gets very high, that fact becomes clear and monetarist models drive interest rate models right out of existence. During the post-WWI era even Keynes and Wicksell started using monetarist models.

    2. I agree that the Fed has all sorts of schemes to affect credit flows, which have nothing to do with the MB. That’s why they have failed, that’s why NGDP is plunging. It isn’t just me who thinks this, James Hamilton (a very good mainstream macro guy) also thinks the credit approach has been ineffective. They need to get back to their job, which is monetary policy. Isn’t it odd that NGDP started plunging rapidly right about the time they started playing around with trying all sorts of new, untested, policies to influence credit markets? Even worse, the lack of an effective forward-looking monetary policy has made the financial crisis much worse. It’s not just that their bailouts and balance sheet adjustments have failed to address the issue, they have made the problem much worse.

    3. Your argument that banks could get a higher yield with longer bonds is a complete nonstarter. The yield curve reflects market expectations. Nobody who holds 3-month T-bills has a gun pointing at their head, whether it is banks or nonbanks, it is because the yield/liquidity combination suits them better than any other asset. And remember, reserves are even more liquid that T-bills. So if T-bills yielding .20 are in equilibrium with T-bonds yielding 2.5% over ten years, then, ipso facto, reserves yielding .25 are superior to long bonds. And don’t forget that with a rising yield curve the expected yield over 3 months on longer term assets is much lower than the average yield to maturity.

    4. The comment about avoiding “punitive” policies on reserves presumably is addressed to my interest penalty idea. As I have indicated in many places, because of the marginal/inframarginal distinction, the plan can be impemented with zero impact on bank profits (and still constrain reserve demand at the margin.) Simply pay a high interest rate on inframarginal holdings.

    5. I have seen the Bernanke quote, but it doesn’t answer any of my questions. Yes, it is to set a floor on the Fed funds rate. As Robert Hall pointed out earlier, that is a “confession” of the Fed’s contractionary intent. Why do they have this intent?

  23. Gravatar of JKH JKH
    20. March 2009 at 12:26


    Thanks much for your detailed response.

    This is a Grand Canyon of a difference in views of the world.

    Great blog by the way. I hope to learn some things here.

    You can no doubt tell I’m not an economist.

    (I developed my views largely as a result of managing the Bank of Canada reserve account of one of Canada’s major banks in the early 80’s. Interest rate levels were slightly different then. But my personal lesson learned at that time was that the policy rate is the center of the monetary policy universe; not the base.)

  24. Gravatar of Jon Jon
    20. March 2009 at 13:54

    “My thought was that there aren’t different types of reserves. You wouldn’t say Bank One has “type A reserves” injected through swaps, whereas Bank Two has “type B reserves” injected through bond purchases.”

    Is your argument that money held in reserve is money sterilized so it hardly matters if its held on account for A or B?

    “I understand that where the money is located has implication for velocity, but not how the currency was injected.”

    My point was that the swaps were specifically intended for foreign banks. Consequently the method implies a locality.

    From there talking in aggregates obscures the further distinctions. For instance, I do think it makes a difference who “owns” the money because that affects the trace and velocity of the funds through the economy. If banks get the funds directly, they have to find willing borrowers. If the public gets the funds and hold them at banks, then banks have to find willing borrowers, but the public could also choose to liquidate those funds into cash and begin circulating the funds.

    Put another way, the efficient markets hypothesis works when the public gets the funds but not when the Fed chooses a priori to give the funds to the banks directly.

    This is but one of the problems wherein the Fed is picking particular markets in which to intervene.

  25. Gravatar of ssumner ssumner
    21. March 2009 at 06:14

    Jon, I don’t understand your argument at all. Banks decide how much reserves to hold (above the res. req.) If they don’t want to hold reserves, they can simply buy assets. And the funds that went to European banks would only stay there if they wanted to hold them, and would have gone there quickly in any case (if that were true) even if initially injected into America. What do you mean by the phrase “have to find willing borrowers.”

    JKH, Interesting observation. Actually I did think you were an economist, as you seem to approach issues from a very pragmatic Keynesian direction. I can see that you know more about the real world than me, but perhaps less abstract theory. So let me try another way of giving you the intuition behind the “monetary” or “non-interest rate” approach to central banking.

    Suppose the Fed or BOC lent unlimited funds through their discount window to any qualified bank. And (although this is not true) suppose the central bank allowed banks to profit from arbitrage opportunities with these funds. Also suppose the discount rate is currently 4% and the T-bill yield is 3.7%. No arbitrage opportunities.

    Now assume a boom or rising inflation expectations start to push free market interest rates up, say to 4.4%. The discount rate stays at 4%. (Again, this is not how real world central banks behave, but I’ll come back to that.) This might lead to explosive growth in the monetary base. You might think that the base growth would drive free market rates back down to 4%, eliminating arbitrage opportunities. But the problem is that the base growth would provide further fuel to the fire, raise growth and/or inflation expectations even further, which further raises the Wicksellian equilibrium rate.

    This creates an indeterminacy, which can spiral into hyperinflation. Not just in my view, but (unless I am mistaken) also the view of the smartest Keynesian economists like Michael Woodford. So why don’t we see this happen? One reason is that central banks may discourage discount loans. But that isn’t the only reason, as you have the same problem from interest rate pegging via OMOs. If your target rate becomes even slightly too low, you may need to pump more and more money into the economy to hold the peg, which creates hyperinflation. So again, why don’t we see this. The answer is that central banks know about this problem, and if nominal spending starts rising too fast, central banks raise interest rates sharply (The Taylor rule, etc.) to eliminate these arbitrage opportunities.

    Now consider my previous example where policy spiraled out of control and led to hyperinflation. What did the two policy indicators show? The discount rate stayed fixed at 4%, while the monetary base spiraled upward by zillions of percent. Which indicator better reflected the stance of monetary policy? Woodford would say that if you look close, the interest rate did signal a highly expansionary policy, as it fell increasing far below the Wicksellian equilibrium rate. I simply find the monetary base an easier way to visualize the policy stance.

    Furthermore, when nominal rates fall to zero all a Keynesian can say is hold them there and hope for the best. A non-Keynesian can say pump in money like crazy, to generate inflation expectations. Buy assets that are not close substitutes for reserves. Krugman would say that’s fiscal policy, as you might lose money when the crisis was over and the money was withdrawn. But if the central bank has inside information, it should buy assets that will appreciate, and if it doesn’t have inside information, there is no expected loss according to the EMH.

  26. Gravatar of Jon Jon
    21. March 2009 at 06:57

    Scott: Foreign banks do not have access to the Fed Funds market. They trade in the eurodollar market (as do domestic banks), but the whole reason for the swap lines was that the eurodollar market ceased to function or rather it functioned ‘too well’. Thus the unusual libor-ff spread that caused the swaps to be created.

    See also for other effects:

  27. Gravatar of ssumner ssumner
    21. March 2009 at 12:16

    Jon, Sorry to show my ignorance on this, but wouldn’t that mean they don’t hold U.S. dollars as part of their reserves? And if so, any swaps you describe which added to their reserves would not have added to the U.S. dollar monetary base. Is that right? But an earlier comment by someone (you of JKH) said that these swaps did add to the U.S. monetary base, if my memory is correct.

  28. Gravatar of Jon Jon
    21. March 2009 at 13:46

    They are accounted for in the base, thanks for arguing through my mistake. The Fed does in fact believe that most of the funds end up back in the reserves of US depository institutions. I wrongly believed that were tracked via the custodial accounts at the Fed directly.

    “As a result, the extension of liquidity through swap lines has resulted not in an increase in foreign official deposits but rather in an increase in reserve balances held by depository institutions.”

  29. Gravatar of JKH JKH
    22. March 2009 at 03:50


    I’m guessing I must read/sound like a broken record, but here is how I would approach the specific example you’ve provided. I’ve distinguished between the “normal case” where interest rates haven’t hit the zero bound and the abnormal case, where we are now. Your example of course is a normal case example in this sense, so most of my comments are on that. I’ve only summarized the abnormal case in the briefest way.

    (It seems to me more generally that a lot of debate confusion these days occurs because of an unfortunate entanglement of normal and abnormal circumstances in the premise, which leads to weird ideological discussions – e.g. using Austrian medicine to treat a patient whose life history is non-Austrian.)


    This is how I would expect money market conditions to unfold in relation to your example:

    First, the overnight Fed funds rate market is highly liquid and highly competitive. The Fed determines the aggregate supply of clearing balances the banks hold at the Fed. So it determines the effective fed funds rate in the sense that it can always take OMO action to steer the effective rate back toward the target rate. If the target rate is 4 per cent, and funds are trading at 3.75 per cent, the Fed will nudge the rate back toward 4 through marginal tightening. And the order of magnitude in “marginal” is tiny, relative to the size of the financial system and even to the size of the total monetary base. Essentially, the Fed fine tunes the path of the effective fed funds rate by manipulating the level of excess reserves (in the form of excess clearing balances). This is my contention, which may or may not require a suspension of disbelief.

    Second, the issue of discount window borrowing is a technical detail which shouldn’t blur the overall analysis. Other things equal, if a particular bank borrows from the discount window, the Fed may or may not offset the effect with a countervailing system reserve drain. It all depends on money market conditions as signalled by the effective fed rate and whether or not that rate is behaving the way it should relative to the fed’s target rate. In other words, the discount window is just another “trade” that the Fed takes into consideration in assessing and reacting to money market conditions as reflected in the trading level for the fed funds rate. It just so happens this particular trade affects the aggregate level of system clearing balances, and the Fed needs to consider potential offset in order to keep the funds rate on path.

    Third, the normal course is for the Fed to steer the effective rate continuously toward the target rate. It is a ground hog day operation. Each day, come into work, and do what you need to do to get the rate back toward the target rate. Respond to all market activity accordingly and consistently.

    Fourth, I’m assuming the target fed funds rate in your example is 4 per cent.

    Fifth, I’m assuming the Fed is successful in its ground hog day mission. It continuously steers the path of the effective fed funds rate back toward 4 per cent.

    Sixth, the meaning of your example depends on whether your rates of 3.7 per cent and 4.4 per cent are overnight rates or term rates. I’m going to assume they’re not overnight rates. Otherwise, that contradicts my ground hog day operation, as the Fed can easily take corrective action to steer the rate back to target. Therefore, I’m going to assume they’re term rates.

    So what we have is a yield curve. I suspect you’ve written entire books on the yield curve, so I don’t need to go into detail. Except that the fed funds rate is a policy rate, and all other rates are market rates. So your 4.4 per cent is a market determined term rate.

    Backtracking a bit, given my normal times assumption, and the nature of the ground hog, there’s no reason to assume anything particularly extraordinary about what’s happening to the monetary base. You say that banks will be loading up at 4.4 per cent, but my reading of the information suggests this is a yield curve or interest rate “gapping” play. Provided that the Fed is happy with funds at 4 per cent and generally happy otherwise, there’s no reason for them to be particularly unhappy about the state of the yield curve or this particular gapping play.

    But my most critical observation here would be this:

    If banks really want to borrow from the discount window at 4 per cent to earn yield curve profit at 4.4 per cent, OK. Presumably they have no problem with what is usually the “stigma” effect of doing so. But there’s no reason why the Fed wouldn’t offset discount window borrowing with systemic reserve drains, such that the overall level of reserves would not remain excessive, and such that the fed funds rate would maintain its path around 4 per cent, as opposed to be driven lower by the gross excess reserves injected with discount window borrowing. In other words, the Fed keeps its eye on the ball vis a vis the 4 per cent funds rate. And therefore there’s no reason to think that the monetary base would undergo any sort of significant NET expansion merely due to discount window borrowing. So my most relevant conclusion here contradicts your assumption of monetary base growth as a result of my interpretation of how the Fed would actually operate here. So there’s no inflationary implication due to the monetary base behaviour per se, because the Fed wouldn’t allow the monetary base to behave this way in the context of your problem, even with unlimited discount window borrowing. Therefore, the market would not form its inflation expectations from a growth in the monetary base, because there wouldn’t necessarily be any (of substance). It would form its inflation expectations from just about everything else that’s going on though – credit conditions, actual inflation, etc. etc. – in other words, the same stuff that the Fed is using to form its own expectations and run policy. To the degree that a positive yield curve reflects market expectations for future inflation or future policy rate increases, so be it. The Fed also uses that information in making its decision on the actual path of the funds rate going forward. The yield curve is a communication device from the market to the Fed about policy evaluation and prospects in that sense.

    On your further point, if the target rate becomes “too low” according to the market’s vote on the matter that would indeed suggest to me that the yield curve should steepen. This would be consistent with a market vote of non-confidence in Fed policy, perhaps due to inflation expectations. In the worst of inflationary expectations (such as the early 80’s), the market’s violent yield curve action might well influence the Fed to raise the target rate pronto rather than wait any longer. But that’s still the Fed’s decision. The market never “forces” the hand of the Fed; not really. Given even a nominal supply of excess reserves in the system, it’s unlikely and even virtually impossible that the Fed would not be able to induce overnight fed funds to trade at around the target rate, even if the market expected a hike the next day. On the other hand, there are times when the Fed resists the yield curve signal given by the market and keeps the target rate where it is, and it simply provides enough excess reserves to the system in order to do this. Banks would rather earn overnight interest at x per cent (normal times) than earn no interest. The fact that they expect to earn x + 50 basis points tomorrow isn’t a reason not to earn X today.

    So in conclusion I would say that you won’t see the monetary base spiralling out of control in the way you describe (again these are normal times), simply because it contradicts my own view of how the Fed controls very tightly the pivotal portion of the monetary base, which is clearing balances. Whenever you see the monetary base clearing balances doing something unusual (again normal time) it’s not because the Fed has “lost control” of the base; it’s because the Fed has permitted that particular base behaviour because its consistent with its short term management of the effective fed funds rate level.

    ABNORMAL TIMES (e.g. now)

    Quite a more elaborate discussion, but from my perspective would still be consistent with the view that the Fed focuses primarily on the funds rate to implement policy. It’s exposed to uncharted territory on its exit strategy from its current extraordinary policies, which will be a return trip from the heights of an abnormal balance sheet back to a more earthbound and normal balance sheet. But the return trip will have some path of policy rate tightening at its core. Seat belts for the return trip includes the payment of interest on excess reserves, and the management of the asset liability mix and term structure of its abnormal portfolio, including a gradual maturing/selling/siphoning off of excess assets and excess clearing balance reserves, with the objective of a smooth deceleration and re-entry.

  30. Gravatar of JKH JKH
    22. March 2009 at 04:17

    Re the foreign currency swaps:

    There’s some underlying accounting detail that doesn’t matter to the interpretation of the substance here (i.e. accounting for the treatment of foreign currency as “collateral” for the primary US dollar leg, and the fact that the currency “swap” exposure is fully hedged by forward FX contracts). But here’s my interpretation of the substance.

    The primary leg of the swap from the Fed’s perspective consists of extending US dollar credit to foreign central banks. This is a dollar asset on the Fed’s balance sheet. The Fed used to include this in the “other asset” classification, but I notice it’s now highlighted more explicitly as its own category.

    These dollars are extended as further credit from foreign central banks to individual banks in their systems.

    Some major foreign banks have direct clearing accounts with the Fed. So the dollars they receive will clear to a reserve credit at the Fed. Others have indirect clearing arrangements with banks that do clear at the Fed. Their dollars will end up as deposits in their clearing banks, and the clearing banks will receive credit at the Fed.

    Either way, the Fed’s dollar swaps (asset) result in a corresponding credit in clearing accounts at the Fed (liability). The result is that new Fed dollar credit creation in the form of swaps has resulted in new clearing account monetary base creation. Although the trip is geographically circuitous, the monetary base effect is no different than that resulting from the Fed’s domestic facilities. The Fed’s substantive position with respect to this dollar swap credit extension is that of global central banker for the US dollar globally.

    (This all assumes other things equal. Of course, the Fed combined its swap activity in the early stages with “sterilization” of some excess clearing balances through Treasury bill financing, with treasury depositing its own balances at the Fed, thereby draining bank clearing balances. This started before the Fed began to pay interest on reserves. And a lot of other stuff goes on with the balance sheet, all of which has the potential to adjust clearing balance levels, depending on the Fed’s full set of choices and decisions in its asset liability management.)

  31. Gravatar of JKH JKH
    22. March 2009 at 05:18


    Just to summarize my interpretation of the difference in Fed policy between normal (positive fed funds rate) and abnormal (zero bound):

    The difference is in the financial intermediation function of the Fed. The Fed has expanded and increased its credit function (qualitative or credit easing) as it has approached the zero bound.

    The Fed’s intervention was due to the failure of the private sector banking system to perform the same credit function adequately. The kernel of private system failure was the reluctance of banks to lend to each other, on the basis of their own perceived/real credit risk. The credit risk problem was in turn a function of a perceived/real capital adequacy problem. The entire banking system seized due a bank solvency perception/reality problem. All of the Fed’s credit activities and asset expansion reflect this private system failure.

    Therefore, any interpretation of systemic illiquidity failure must acknowledge the core problem of insolvency perception and/or risk, of which liquidity attributes are merely the symptom.

    In this context, the increase in the monetary base is a choice that the Fed has made in order to complete the financial intermediation of its credit activities. It does not reflect a primary objective of monetary base expansion. It is a consequence. The payment of interest on reserves is designed to neutralize the unintended consequence – which would be the loss of control over the fed funds rate at such future point as when the Fed begins to tighten the policy rate even though its extraordinary intermediation activity has not been fully unwound.

    The expansion of the monetary base therefore is primarily intended as a Fed funding mechanism – not a banking system liquidity motivation mechanism. The latter impetus must come from a restoration of credit conditions and capital adequacy for the banking system.

    Banks don’t need excess clearing balance reserves in order to extend credit and take on non-zero credit risk (i.e. not treasury bills). They need capital and some general sense of confidence in the economic outlook. Under such conditions, any bank with reasonable credit status itself can fund newly extended credit through liability management, since the system in aggregate creates money from credit. Banks don’t need clearing balances to fund expansion. Aggregate credit funds aggregate money; not vice versa. So the idea that the Fed is planting excess reserves in the system now in an attempt to jump start credit lending is not really the case. It is attempting to jump start credit lending by pump priming system credit risk taking through its own extraordinary credit activities. Excess reserves are a residual effect of this strategy rather than the strategy itself.

  32. Gravatar of Jon Jon
    22. March 2009 at 09:57

    JKH: You paint with too large a brush. The two programs contributing most of the reserves are the TAF and CPFF, both of these were created explicitly with the notion that there was a liquidity crisis. So the system was explicitly flooded with liquidity and the reserves are not a side-effect of other pump-priming activities.

    Critically also, this liquidity did not arise from deposits but represents direct auctions of money by Fed to depository institutions. Thus, there is no risk to a bank-run. To understand this policy, you need only look back over 2008. The discount window is heavily stigmatized and fears of bank-runs were widespread and well founded. This is precisely why the FDIC insurance limit was raised, and why the CPFF has primarily been used to liquefy bank paper.

  33. Gravatar of Alex Alex
    22. March 2009 at 10:45


    Nominal liabilities alone don’t tell the whole story. If increases in debt are matched with higher taxes in the future then they don’t have to be inflationary. I think that the US will not be able to tax and pay all of the new debt so in the end it will just have to inflate away. BTW you can already see the effect on the dollar, gold, oil and TIPS.


  34. Gravatar of Alex Alex
    22. March 2009 at 10:52

    Ohhh I forgot to mention, that my figure includes only the data until February 2009, so all the measures taken in March are not shown there.

  35. Gravatar of JKH JKH
    22. March 2009 at 11:14


    I plead guilty to the large brush charge, but only in the mildest of ways. I’ve emphasized solvency over liquidity, primarily to make a point about the reason for the existence of large clearing balances. But I didn’t say there wasn’t a liquidity crisis. I said banks wouldn’t lend to each other because of solvency fears. While TAF is an extension of the normal Fed liquidity provision function, the problem it’s designed to address has solvency concerns at its roots – so with all of the extraordinary Fed credit facilities. They all improve system liquidity from what it otherwise would be in these circumstances, but not because they create bank clearing balances in the process. The mere fact that there are such large clearing balances now and the banking system is still not functioning properly means there’s something larger than liquidity at play. The idea of a “liquidity trap” means there’s something more at play. And the ongoing functioning of so many credit facilities means you can’t attribute one or the other to the base creation; it’s all one giant machine of credit extension and base creation. All of these programs are necessary due to the overall solvency status of the system. I don’t know about the CPFF, but I thought it was targeted toward problems emanating from the shadow bank or non bank system. But again, the core purpose was not to create additional bank clearing balances. Nor with the FDIC limit increase. Both are liquidity issues, but not for reasons related to the size of bank clearing balances, and both liquidity issues have credit risk issues at their core.

  36. Gravatar of Jon Jon
    22. March 2009 at 14:38

    “And the ongoing functioning of so many credit facilities means you can’t attribute one or the other to the base creation; it’s all one giant machine of credit extension and base creation.”

    But I can by notional value. I can also use ‘logic’, the TAF is a direct auction of money it has no immediate impact on any credit market itself. “By allowing the Federal Reserve to inject term funds through a broader range of counterparties and against a broader range of collateral than open market operations, this facility could help ensure that liquidity provisions can be disseminated efficiently even when the unsecured interbank markets are under stress.”

    I agree that there was a solvency problem, but the Fed was pretty clear at denying the solvency problem. The programs were explicitly designed to address liquidity concerns. So much as solvency was at the root, no real progress has been made. That’s precisely the problem. Arguing that there was a grand scheme and the reserve balances were side-effect is disconnected to the facts. Large reserve balances were the goal.

    I agree, however, that it does not follow that the Fed expected those reserve balances to encourage lending. As far as I can ascertain, they took a neo-keynesian position that the price signal of lowering short-rates would do that work. Sadly for them, us, and the neo-keynesian position events have falsified that theory of transmission.

    Monetarist theory predicts this policy failure as a direct consequence of there having been no significant effect on the quantity of money in circulation.

  37. Gravatar of JKH JKH
    22. March 2009 at 21:45

    Sorry, but you are wrong on the TAF.

    The Fed created TAF on December 12, 2007.

    Excess reserve balances did not begin to increase noticeably until October, 2008. The data is in the time series summaries of the weekly Fed releases.

    This means that the Fed sterilized the full specific TAF effect for nearly a year, until the banking crisis reached an entirely different level with additional credit related Fed responses.

    Therefore, the objective of TAF was certainly not to increase the monetary base.

    The Fed was entirely guided by ensuring the level of excess reserves that in turn ensured an effective fed funds rate that tracked to target.

    And this continued until the Fed couldn’t introduce additional credit programs without increasing the monetary base as a source of financing. And that’s why they started to pay interest on reserves.

    The very fact they pay interest on reserves essentially proves that they have no real interest in inflating the monetary base per se. Expansion of the monetary base is primarily a funding mechanism for the Fed; not a motivation mechanism for the banks that hold the excess clearing balances that have inflated the monetary base.

  38. Gravatar of Jon Jon
    22. March 2009 at 22:19


    As you ought to know, there is a difference between deposits and other liabilities. The Fed did drain funds by disgorging itself of TBills. Those sales replaced deposits with tbills. The Fed then auction the money to depository institutions. Now the banks had more liabilities to the Fed and fewer liabilities to the public. And in particular, those banks needing the liquidity could get it from auction, whereas those banks that did not would not.

    The TAF was a liquidity program. Rising excess balances are not necessary for this to be so.

    “The Fed was entirely guided by ensuring the level of excess reserves that in turn ensured an effective fed funds rate that tracked to target.”

    No one here disputes that the Fed uses its FF target to regulate its OMO. No one here has claimed that the Fed uses a money aggregate target. The disconnect is in explaining how the Fed’s policies effect the economy and the subsidiary function of the Fed’s special programs.

  39. Gravatar of JKH JKH
    23. March 2009 at 03:24

    You contradicted yourself.

    Your comment 22 14:38:

    “Large reserve balances were the goal.”

    Your comment 22 22:19:

    “Rising excess balances are not necessary for this to be so.”

    On the other hand, I never denied that TAF is a liquidity program.

    My comment 22 11:14:

    “While TAF is an extension of the normal Fed liquidity provision function …”

    You can characterize the TAF as you wish. In fact you can characterize it as the Fed has done, and as I did already above – as a liquidity program. The subject I’ve been addressing is the behaviour and purpose of the monetary base in aggregate from the Fed’s perspective, and I’ve been consistent in focusing on that aspect with whoever might be interested in engaging. I don’t need to be told to change the subject if I’m interesting in pursuing a mischaracterization of the topic that I’m primarily interested in as a prerequisite to how that topic connects to wider economic issues.

    Thanks for the discussion.

  40. Gravatar of ssumner ssumner
    23. March 2009 at 05:56

    I won’t comment on the debate between Jon and JKH, because I don’t know enough about these programs. On other issues:

    JKH, I think you misunderstand my hyperinflation example. If the Wickellsian rate rises to 4.4%, I don’t mean that is the market rate, I mean is the the potential market (overnight) rate that would insure price stability. To prevent the fed funds market rate from rising to 4.4%, the Fed must inject base money to hold down the ff rate. This is standard theory. But if they do this, inflation expectations will rise further, pushing the Wicksellian equilibrium rate even higher. This starts a vicious spiral that will definitely lead to hyperinflation if the Fed doesn’t allow the fed funds rate to rise above 4.0%. There is no real controversy over this point among monetary theorists.
    You correctly note that it seems like the Fed is always able to move the short term rate where they want to. Yes, but only because the adopt Taylor Rule-type policies. That is, when inflation starts to rise they raise rates by even more than inflation rose. That raises the real rate, and slows down inflation. If they didn’t do that, then we would get hyperinflation. If you go back and read my example with all this in mind (and it is not controversial) it might make more sense.

    JKH#2, Even if you are right about the need for qualitative easing, I would make two points.

    1. I doesn’t eliminate the need for QE.
    2. If the Fed did its job the banking system never would have malfunctioned in the first place. The primary cause of the malfunction was not subprime loans, it was tight money that caused NGDP to plummet.

    You might also want to take a look at my earlier post “A short course on money” if you want to better understand my perspective. Fed policy doesn’t even require the existence of a banking system, or even “interest rates”. You could simply have a Fed dealing with the public, buying some asset for OMOs, and controlling the money supply that way. That would still control the price level. The banks do not play any sort of crucial role in the process–they just appear to.

    Alex, I disagree. TIPS are by far the best of the inflation indicators you mention, and they show just over 1% expected inflation over the next decade. We need higher inflation expectations, especially near term.

  41. Gravatar of Alex Alex
    23. March 2009 at 06:34


    Here are the yields on index and non indexed bonds.

    Notice how the expected inflation has been increasing all over 2009 and we know that one the inflationary expectations set in the yield on non index bonds will shoot through the roof. Another thing, my 33% estimate in inflation is not 33% inflation for ever. It is an accumulated 33%. It could be 33% in one year and 0% ever after or 3% higher inflation in perpetuity (actually something that in present value gives 33%). If I was China I would start moving my reserves to TIPS instead of T-bills and notes.

  42. Gravatar of Jon Jon
    23. March 2009 at 07:35


    I leave you with this:

    Read into it what you will.

  43. Gravatar of JKH JKH
    23. March 2009 at 10:58


    Thanks for your patient responses.

    I’ll try to spend some time looking at your various posts.

  44. Gravatar of TheMoneyIllusion » How do we rebuild balance sheets? TheMoneyIllusion » How do we rebuild balance sheets?
    9. June 2009 at 14:43

    […] Speaking of The Economist, an interesting coincidence here and […]

  45. Gravatar of スコット・サムナーのマネー観、マクロ観 by Scott Sumner – 道草 スコット・サムナーのマネー観、マクロ観 by Scott Sumner – 道草
    7. May 2011 at 23:47

    […] マクロの標準的な見方では金融政策および財政政策ツールの組み合わせは、「長く可変のラグ」を伴ってマクロな経済に幅広く作用するとされる。その伝達メカニズムは「ブラックボックス」のようなものの中にあってほとんど分かっていない。つまりまず政策行動がこの謎の箱の中に入り、一年後にマクロ経済的な結果となって飛び出してくる。VARモデルを使ってその効果を見積もることはできる。私はこれは資産市場が非効率であることを暗黙に仮定していると考えるので、そのような物の見方は受け入れられない。インフレーションに織り込まれる長期の効果は直ちにTIPSスプレッドとCPI先物価格に反映されるはずだ。それが正しい政策評価の方法である。   […]

Leave a Reply