Tuesday, August 6, 2013

A Permanent Expansion of the Monetary Base in a World of IOR

In my previous post, I argued that for monetary policy to be effective in a liquidity-trap like slump, the increase in the monetary base has to be seen as permanent. That is, if the central bank signals and the public believes that a large enough portion of the expanded monetary base will be permanent even after a recovery takes holds, then investors will realize the monetary base and treasury will not be near substitutes in the future. Consequently, they will start rebalancing their portfolios today and that, in turn, will spur growth in aggregate demand.

What I did not spell out is how this would unfold if the Fed continues to pay interest on reserves (IOR). The same conclusion holds, but in addition to the Fed permanently maintaining some portion of its expanded balance sheet it would require the central bank not to raise the IOR as other market interest rates rose. Another way of saying this is that as the recovery pushed up the equilibrium or natural interest rate, a decision by the Fed to keep the IOR pegged at 0.25% will result in more inside creation, more spending, and ultimately higher inflation for a given stock of the monetary base. For a more thorough discussion of this point, along with some summary tables, see this earlier post of mine as well as this one by Steve Randy Waldman.


  1. "inside more creation"===>"inside money creation"?

    The idea that a trivial thing like paying interest on reserves can have significant economic effects is implausible.

  2. The monetary base is uninformative. Nobody knows, or cares, how many $100 bills Russian gangsters will demand in the future. The base will adjust. What matters is the Fed's target.

  3. David, both Jared and I have responses back on pragcap:


  4. OK, but according to this description, “as the recovery pushed up the equilibrium or natural interest rate, a decision by the Fed to keep the IOR pegged at 0.25% will result in more inside creation, more spending, and ultimately higher inflation for a given stock of the monetary base,” it’s the rates that are doing the inflationary work, not the size of the base. Couldn’t your description work with any size base? Couldn’t the Fed skip the QE altogether and just promise to keep rates low for a very long time? How would that change your story?

    The banks don’t need the reserves in order to lend. If they need more reserves to clear all the new payments circulating around from the increase in lending, the Fed will provide them as needed at the low (0.25%) target rate (that’s what targeting a rate means to a central bank). In other words, consistent with your story, the increase in the base need only come AFTER the increase in inside money creation and spending. The expansion of the monetary base plays no causal role.

    Even in the Waldman link you posted, he says “They [market monetarists] may need to rethink stories that place the quantity of base money (as distinct from debt) at the indispensible [sic] heart of macro policy.”

    1. That quote from Waldman was one that stuck out to me as well. Thanks for bringing that up.

    2. I am not following you Jared. Whether the monetary base comes after or before, the inside money creation is conditional on it being there. If there is no monetary base expansion to support the 0.25% IOR then it goes up and there is no inside money creation.

    3. Anonymous,

      You're technically correct. But I’m saying that a central bank that targets an interest rate is already committing itself to providing the needed reserves. My criticism of David is that, while he's also positing a central bank that targets an interest rate (IOR=0.25%), he fails to recognize that this alone ensures the reserves required to facilitate loan creation at any level. By failing to recognize this feature of the system, he claims that the central bank must increase the monetary base BEFORE the increase in lending/spending occurs. I maintain that increasing the monetary base BEFOREHAND is pointless if the central bank is already targeting a rate.

      Furthermore, David believes that increasing the base will induce inflationary expectations. But this is what Tom and I are having trouble understanding. In his previous post, David thinks spending will increase due to the expectation that the return on money and treasuries will diverge. I tried to show that with the presence of excess reserves, money and treasuries will be near-perfect substitutes for a very long time, if not permanently. In this post, David seems to reframe the cause of the inflationary expectations as the promise of future low interest rates. But if that’s the case, increasing the base seems superfluous.

      I hope that helps.

    4. Jared, sorry for the delay been traveling.

      Regarding your first criticism, everyone who makes the claim about the importance of permanent monetary base increases recognizes that pegging an interest rate means the central bank accommodates short-run changes in base demand. My point is that the CB commits to a permanent increase in the monetary base. How it chronologically unfolds is besides the point. Yes, in an interest rate targeting regime the CB accommodates demand. In monetary base targeting regime the increase is explicit.

      I have never argued (or meant to argue) that the monetary base has to increase first. What I have argued is that the CB promises first to allow the monetary base expansion to occur or if it has already occurred (like now) that it will not be reversed. As noted earlier, there can be no increase in inside money if this monetary base expansion does not occur.

      I think part of problem we are having here is that you view monetary policy solely as interest rates. I view interest rates only as intermediate target or indicator variables, not the actual instrument of monetary policy. Along these same lines, I suspect you view all changes in the monetary base as endogenous. I only view short-run changes for a given interest rate target as endogenous. Over longer horizons the Fed is changing interest rates according to something like a Taylor Rule. These policy changes in target interest rates mean exogenous changes in the monetary base. (Actually, it is the parameters of the Taylor Rule that are truly exogenous, but that is best save for another post).

    5. David, I still don't see how the short term Tsy debt rate can rise above the IOR rate while excess reserve stocks are > 0.

      Is cash important for this? That link you sent presented a cashless society. Let's keep that assumption for now.

      Is it the reserve requirements? Are those important? It certainly seems like that's what you're saying with this statement "there can be no increase in inside money if this monetary base expansion does not occur."

      If reserve requirements are 0% though (like in Canada) this doesn't seem so important does it?

      So let's say we have our cashless Canada. Can short term Tsy rates there increase above IOR with ER > 0?

      I added one other assumption (one commercial bank) to make the accounting easier. You can think of it as just aggregating all the commercial banks together. In such a system it's clear the bank can ONLY use its reserves to buy Tsy at auction from Tsy, or use it in Fed OMS (if the Fed was conducting OMS which it's not). EVERYTHING else must be purchased by the bank by crediting privately held bank deposits (i.e. creating inside money). How do short term rates rise above IOR in such a scenario? Won't the bank buy them all up during the Tsy auction? The reserves they use to do so being immediately replenished once Tsy spends the proceeds.

      Have these assumptions gone too far? Which ones? Are non-zero reserve requirements vital to the importance of permanent monetary base increases? Is cash? Is having more than one commercial bank? Why?

      It seems to me that ER > 0 ensures that short term rates cannot rise above IOR even with cash and with reserve requirements > 0%. Only when ER is drained do I see that as a possibility. They can be drained by the Fed doing OMS, by Tsy collecting a surplus, by a huge increase (by 1/(reserve requirements)) in bank deposits (via the bank's buying LOTS of stuff: loans & whatnot) or through the public demanding more cash. Until then won't the banks always trade their reserves to Tsy for debt so they can temporarily be rid of them? What am I missing?

    6. BTW, I made essentially the same point with "bart" the committed hyperinflationist here:

      He responded with the "rehypothecation" and "collateralization" as you can see. He loses me on that one. But do you take a similar position? Is that important?

      (Ha!.. just noticed that Cullen is essentially accusing bart of being a conspiracy theorist for that one.. it does have a ZeroHedge kind of sound to it.. but I don't really understand those concepts well enough to know for sure).

    7. David, sorry for the firehose of replies here, but I think I get what your saying here:

      "I only view short-run changes for a given interest rate target as endogenous."

      I even did a very very amateurish post on that myself (trying to translate it into the engineering world that I'm more comfortable with):

      It's a point I've seen both Nick Rowe and Scott Sumner make:

      I don't necessarily have a problem with that view. But with ER > 0, wouldn't you just do that by setting the IOR rate to the target rate (according to Taylor's Rule)? How does that point about short term endogeneity argue against this?

    8. Oh, and David Glasner too:

    9. Ah!!!! I found this reply very helpful, David. Thanks! But our discussion began with this quote from Friedman, ““Oh well, we’ve got the interest rate down to zero; what more can we do?” It’s very simple. They can buy long-term government securities, and they can keep buying them and providing high-powered money until the high powered money starts getting the economy in an expansion.” This certainly sounds like a causal (and temporal) story running from CB asset purchases (injections of high-power money) to economic expansion. But it seems as if you’re now saying that the central bank does not really need to conduct asset purchases to increase the monetary base, it just needs to credibly commit to a permanent increase in the monetary base, which could FOLLOW the inside money creation. Is your view different from Friedman’s?

      And I’m very glad you brought up the short-run vs long run/endogenous vs exogenous issue. I’ve been struggling with that since Nick’s post that Tom linked to. So, we all agree that in the short run (6 week intervals?) the CB is committed to a target rate and will provide all the reserves needed to accommodate private bank lending. This change to the base we call endogenous. But then what? Let’s say that during that 6 week period bank lending soared and, in turn, the base soared. The CB then sees inflation on the horizon. What does it do? My answer is that it will announce a rate hike. And for the next 6 weeks it will do the same thing that it had done previously, only at a different (higher) rate: adjust the base to accommodate private bank lending. With the higher rate, it is very likely that lending will slow down (more people are priced out of the market). This could lead to three outcomes: the CB might only need to add to the base slightly; or 2) if lending stays flat, the CB would not have to adjust the base at all; or 3) if people start paying off outstanding loans faster than new loans are extended, the CB would have to conduct OMS’s and drain some reserves, shrinking the base. But, again, this is no different than the prior 6 weeks. In both periods, the CB accommodates endogenous changes to the money supply at a targeted interest rate. Never (not since the early 80’s anyway) will the CB target the actual size or growth rate of any monetary aggregate. The only exogenous variables are the interest rate and the inflation target. If you consider the short-run endogenous, you should, in my opinion, consider the long-term endogenous. Moreover, I don’t think it’s about short-run vs long-run at all. To steal from Scott Fullwiler, it’s about tactics vs strategy.

      David, I’m very interested to see how you would describe this process.

    10. "David, I’m very interested to see how you would describe this process. "

      Me too! ... but Jared, do you agree that David still has not addressed how short term rates get above the IOR rate when ER > 0? If he did and satisfied you, I missed it. :(

      I'm actually very confused about what he's saying. I too would like to know if his view is different than that Friedman quote you provide. I really don't see the point of the CB forcing ER > 0 given your line of reasoning. In fact I'm more convinced now that it's NOT necessary. But I'm having a hard time squaring David's statements here w/ his previous statements.

      Regarding Japan, they've gone well beyond "traditional" asset purchases though... they're explicitly targeting a value for their stock market index, right? Doesn't that veer into very different territory? As Cullen says, when the Fed starts paying people for bags of dirt, how's that different from fiscal stimulus? Sumner would say that it doesn't end up on Tsy's balance sheet... that's the difference. OK, but so what?

    11. Jared,

      Let me respond by illustrating with an extreme case. Consider the implicit Taylor rule for Zimbabwe during its hyperinflation period and the Taylor Rule for the ECB now. The parameters on the inflation term (which measure to degree to which policymakers respond to changes in inflation)are very different: 0 in Zimbabwe and very high at ECB. These different inflation parameters are an exogenous policy choice and ultimately determined the very different paths of the monetary base for these two countries. (If you don't like Zimbabwe, then compare 1970s US high inflation period to current ECB practice.)

      Closer to home, when the Fed changes its target interest rate it implies either (1) an discontinuous shift in the supply of bank reserves which is exogenous or (2) different degree of policy accommodation to changes in demand for bank reserves. With better communication and signalling, the second option is the more likely outcome today. But even this outcome can be seen an an exogenous change. The Fed has exogenously chosen to change the intensity of its policy accommodation. This change in intensity is a policy choice that will shape how much monetary base there will be.

      Again, this comes back to what does the Fed really have control over? It is the monetary base. The federal funds rate is simply an intermediate, indicator variable that helps us figure out how much monetary base is needed to hit a particular inflation/NGDP target.

    12. Tom,

      I'm still with you. I still don't know how he thinks short term rates rise above IOR when ER>0. But I dropped it because in this post where he discusses a world in which IOR>0, he doesn't mention the divergence of rates at all (unless he believes that T-bills automatically take on the value of the natural interest rate. But that's pure mysticism!). In this post and subsequent comments, he instead stresses the central bank's PROMISE to permanently increase the monetary base, but also maintains that the increase in the monetary base need not occur before the lending or spending occurs. To me, this sounds like he's denying the efficacy of QE-style asset purchases, which, at first, I thought was in clear contradiction with his Friedman post. But now after re-reading that post, I think David is being consistent.

      I originally read him as saying that QE-style asset purchases were necessary for generating increased demand. But now that he has very clearly stated that the monetary base need not increase first, I interpret that Friedman post as saying that IF QE-style asset purchases have been conducted, THEN the only way for them to be inflationary is for the CB to promise they will be permanent. He also adds that the CB must refuse to raise IOR once spending picks up. So the necessary conditions for generating an increase in demand are 1) a promise that any monetary base expansion will be permanent; 2) actually expanding the base to accommodate loan growth (not necessarily before), and 3) not raising IOR. QE is NOT necessary. Is this interpretation accurate, David? If so, do you also believe these conditions are SUFFICIENT for increasing demand?

      Tom, I guess I've let him off the hook about the divergence between IOR and treasuries because I no longer think it essential to his theory. I'm not saying I agree with him, but I think I'm progressing towards understanding where he's coming from. We'll see from his response.

    13. "Again, this comes back to what does the Fed really have control over? It is the monetary base."

      Consider a hypothetical world where privately produced money is a perfect substitute for base money. In that case, monetary policy could not be conducted by OMOs. But you could implement a Taylor Rule just as easily as now, since the Fed doesn't need to change the quantity of base money in order to change the Fed Funds rate.

      The insistence that monetary policy is "really" about the quantity of base money comes from, I guess, Milton Friedman's idea of money supply targeting. Clearly, if the Fed loses control over the money supply, then money supply targeting can't possibly work. That's a strong hint that it's probably not a sound idea to begin with.

    14. Jared, thanks for your response. Say ER > 0, and we fix IOR. Now I can see that inflation COULD increase, and I can see that longer term rates might increase, but that doesn't change that short term rates or the "FFR" as David wrote here:

      continue to be pegged to our now fixed IOR until ER = 0. Do you concur? So I'll take your word on the re-read of the previous post, but this seems to be a definite change from David's previous comments.

      Also, regarding Friedman's idea (from your quote from the previous article) of continuing to pour in "high-powered money" after ER is already > 0: It seems to me the more you pour in, the greater the time period over which the CB is stuck with ER > 0 and thus also stuck with IOR = FFR. It seems like that just takes away a degree of freedom, and the more of it (increasing ER) that's done, the longer this freedom is taken away. So what's the point? Psychology?

      Say the CB really needed to immediately and dramatically raise the FFR for some reason, but ER >> 0. They'd be forced to do that by raising IOR dramatically, wouldn't they? Whereas if ER = 0 the whole time, they could increase FFR w/o changing IOR. Seems like they're just tying their own hands by continuing to make ER >> 0.

      Max, I think the world you refer to might be similar to the one I postulated: No cash, No RRs, one commercial bank. The CB still has to provide reserves for the bank to borrow for Tsy bond auctions and for (anybody's) tax payments to Tsy, so they could set the FFR. But you're right, if that's ALL they want to set, then no OMOs are needed. If they also want to set the yield on a 10-year note, they have to do OMOs. I *think* that they have to do OMOs for the FFR if we allow multiple banks, but I'm not sure because w/ RR = 0, the banks don't have to hold reserves overnight.

    15. Jared & Max: it seems to me the only reason a "monetary base" is needed overnight at all at the banks is because reserve requirements (RR) > 0 and cash. Take those away, and the "base" could disappear from the banks entirely each and every night. So long as the CB promised to provide them as needed, why would they EVER have to have any at the banks overnight? The only overnight "base" needed would be in the Treasury General Account (TGA), and the Fed deposits of GSEs, etc.

    16. Tom, that's right. Ironically, a Friedman-style monetarist has to constantly worry about the free market undermining his monetary order. Not very libertarian!

    17. Even in the absence of regulations banks will still hold some reserves. For example, see the 100 year plus experience of free banking in Scottland where no there were no required reserves

      Banks hold reserves for both precautionary reasons and for and expected withdrawals. U.S. Banks held excess reserves pre-2008

    18. Anonymous, I agree that banks will still hold some reserves. I've seen Nick Rowe claim that about Canada. Thanks for the links... especially the chart. So it looks like it hovered between $1B and $2B from 1984 to 2008. That's NOT a lot though, you have to admit! As payment clearing systems, etc. have become more efficient, and if IOR = 0% (or less!), I'd think banks would be incentivized to try and hold as little as they could get away with.

  5. Anonymous, the context here is that in the comments to David's previous post and on the links I provide above, Jared is essentially arguing that ER > 0 implies FFR = short term Tsy rates = IOR. If that's oversimplified Jared, excuse me. I'm with Jared on that (if that's what he really meant!)... and both of us are confused (at least I still am!) how David can talk of rates of short term Tsy debt rising above IOR while maintaining ER > 0 the whole time.

    So what he's saying here is from that context. To address your comment, why does IOR have to go up? That can be set independent of anything else. Even if the FFR and short term Tsy debt rates climb above 0.25%, the Fed can still keep IOR at 0.25%. In this case (however) I don't see how ER stays > 0.

  6. "...but in addition to the Fed permanently maintaining some portion of its expanded balance sheet it would require the central bank not to raise the IOR as other market interest rates rose."

    Correct me if I'm wrong, but this sounds very much like what New Keynesians (Woodford and Eggertson) are saying with respect to forward guidance, or, alternatively, like Krugman's "credibly promise to be irresponsible".

  7. David,

    If we cut through the jargon, by "different degree of policy accommodation to changes in demand for bank reserves," you mean the Fed will raise or lower the rate at which it provides reserves to private banks, right? So far in our discussion, you have admitted that increasing the monetary base in itself will not raise spending. It’s the expectation that the central bank will accommodate easily even after inflation picks up that will generate spending now. But what “accommodate easily” amounts to in practice is keeping rates (IOR) low. Yes, this must be accompanied by an increase in the monetary base, but that is the OUTCOME of the promise to keep rates low in the face of inflation. If the Fed increases the monetary base first, which it has, but indicates IOR will soon rise, your theory predicts stagnation. But if the Fed were to keep the monetary base at its current level and guarantee that IOR would not rise for, say, 10 years, then your theory should predict increased spending. I think this exemplifies that it’s the rates that matter; the size of the monetary base is passive. The CB will never know “how much monetary base is needed to hit a particular inflation/NGDP target” until it hits it.

    And just as JP posted above, this sound very much like Krugman's "credibly promise to be irresponsible".

  8. "If we cut through the jargon, by "different degree of policy accommodation to changes in demand for bank reserves," you mean the Fed will raise or lower the rate at which it provides reserves to private banks, right?"

    Great point Jared!

    1. "it would require the central bank not to raise the IOR as other market interest rates rose"

      "other" is perhaps a key word there in terms of acknowledging that ER > 0 implies FFR = IOR? ... but it's saying longer term rates (the rest of the yield curve) could rise (assuming IOR did not)?

      Jared, I'm still on board with your line of reasoning there, but perhaps there is an issue which I don't think I've seen addressed fully in these two posts and comments. David writes:

      "then investors will realize the monetary base and treasury will not be near substitutes in the future. Consequently, they will start rebalancing their portfolios today and that, in turn, will spur growth in aggregate demand."

      There's still that bit about "not be near substitutes" which still bothers me, since as you pointed out Jared, it'll be a LONG time before ER = 0 again. It seems the word "other" in David's comment above might hint that FFR is acknowledged to be pegged to IOR? You're letting David off the hook for that, and I get it: that's fine.

      So while I question that expected "near substitutes" description still, the "portfolio rebalancing" part is something that I haven't really considered in my comments up until now. I prefer to think that some of this portfolio rebalancing will be a consequence of the added commercial bank deposits that QE has added. After all QE does make ER > 0, but it also changes non-bank private sector Tsy debt holdings into bank demand deposits (when the non-banks are net sellers to the Fed), right? So the people that sold are going to look to find replacement yield... or at least have other plans for their money (nobody sells unless they have a plan, right?.. it's not like the Fed is FORCING anyone to sell). That's where I've seen others describe the MM transmission mechanism like this: "QE raises prices (and reduces yields) on ALL financial assets, thus causing more investment in real goods and services." Real goods here could be a sports car or a new factory (that promises to provide a greater risk adjusted rate of return than what's available by buying Tsy's). So instead of just tying the Fed's hands by forcing FFR = IOR whenever ER > 0 (as I stated earlier), perhaps more importantly there's this "portfolio rebalancing" happening (hot potato?) which more than makes up for that. Does that make sense to anyone?? :^P