Interest Rate Determination

By L. Randall Wray [via CFEPS]

A few years ago, textbooks had traditionally presented monetary policy as a choice between targeting the quantity of money or the interest rate. It was supposed that control of monetary aggregates could be achieved through control over the quantity of reserves, given a relatively stable “money multiplier”. (Brunner 1968; Balbach 1981) This even led to some real world attempts to hit monetary growth targets—particularly in the US and the UK during the early 1980s. However, the results proved to be so dismal that almost all economists have come to the conclusion that at least in practice, it is not possible to hit money targets. (B. Friedman 1988) These real world results appear to have validated the arguments of those like Goodhart (1989) in the UK and Moore (1988) in the US that central banks have no choice but to set an interest rate target and then accommodate the demand for reserves at that target. Hence, if the central bank can indeed hit a reserve target, it does so only through its decision to raise or lower the interest rate to lower or raise the demand for reserves. Thus, the supply of reserves is best thought of as wholly accommodating the demand, but at the central bank’s interest rate target.

Why does the central bank necessarily accommodate the demand for reserves? There are at least four different answers. In the US, banks are required to hold reserves as a ratio against deposits, according to a fairly complex calculation. In the 1980s, the method used was changed from lagged to contemporaneous reserve accounting on the belief that this would tighten central bank control over loan and deposit expansion. As it turns out, however, both methods result in a backward looking reserve requirement: the reserves that must be held today depend to a greater or lesser degree on deposits held in the fairly distant past. As banks cannot go backward in time, there is nothing they can do about historical deposits. Even if a short settlement period is provided to meet reserve requirements, the required portfolio adjustment could be too great—especially when one considers that many bank assets are not liquid. Hence, in practice, the central bank automatically provides an overdraft—the only question is over the “price”, that is, the discount rate charged on reserves. In many nations, such as Canada and Australia, the promise of an overdraft is explicitly given, hence, there can be no question about central bank accommodation.

A second, less satisfying, answer is often given, which is that the central bank must operate as a lender of last resort, meaning that it provides reserves in order to preserve stability of the financial system. The problem with this explanation is that while it is undoubtedly true, it applies to a different time dimension. The central bank accommodates the demand for reserves day-by-day, even hour-by-hour. It would presumably take some time before refusal to accommodate the demand for reserves would be likely to generate the conditions in which bank runs and financial crises begin to occur. Once these occurred, the central bank would surely enter as a lender of last resort, but this is a different matter from the daily “horizontal” accommodation.

The third explanation is that the central bank accommodates reserve demand in order to ensure an orderly payments system. This might be seen as being closely related to the lender of last resort argument, but I think it can be more plausibly applied to the time frame over which accommodation takes place. Par clearing among banks, and more importantly par clearing with the government, requires that banks have access to reserves for clearing. (Note that deposit insurance ultimately makes the government responsible for check clearing, in any event.)

The final argument is that because the demand for reserves is highly inelastic, and because the private sector cannot increase the supply, the overnight interest rate would be highly unstable without central bank accommodation. Hence, relative stability of overnight rates requires “horizontal” accommodation by the central bank. In practice, empirical evidence of relatively stable overnight interest rates over even very short periods of time supports the belief that the central bank is accommodating horizontally.

We can conclude that the overnight rate is exogenously administered by the central bank. Short-term sovereign debt is a very good substitute asset for overnight reserve lending, hence, its interest rate will closely track the overnight interbank rate. Longer-term sovereign rates will depend on expectations of future short term rates, largely determined by expectations of future monetary policy targets. Thus, we can take those to be mostly controlled by the central bank as well, as it could announce targets far into future and thereby affect the spectrum of rates on sovereign debt.

26 responses to “Interest Rate Determination

  1. A somewhat related question is whether the Fed can target EXCESS reserves. The implicit assumption in most "exit plan" discussions is that the Fed controls those reserves directly and specifically. They injected those reserves; they can withdraw or maintain them as excess.But how do banks view excess reserves? Not as some "monetary aggregate". Individual banks simply view them as cash substitutes (for T-bills). Under this view, today, banks want to hold roughly $700b in cash, and that has not changed despite the impovement in financial conditions.If current bank cash/liquidity preferences are lasting, then a Fed attempt to remove "Excess Reserves" would have little impact on bank cash holdings. In an effort to maintain proportionate cash levels while deposits fell, banks would, in fact, reduce lending and Required Reserves. This would obviously be catastrophic for the money supply.Similarly, if bank cash preferences change, banks would lower credit standards and lend out ER's, and the multiplier effect would be dramatic — imagine even $100b of ER's flowing through to deposits and being re-lent, over and over. How can the Fed control the ER levels in that event? Using the interest rate payment on reserves. However, think — banks may be quick to lend out, repeatedly, that $100b, but the Fed will certainly not be quick to raise the interest rate. Would 1% stop the $100b from being lent? 2%? 3%? Unlikely. And yet NOONE expects 2-3% Fed Funds rates in the next two years.So the Fed has a highly inneffective tool for controling massive moves in the money supply. Does anyone else realize this?

  2. Let's be real, they have no control, and bow to the requests of people that control many more times the financial capital than any of them personally do.

  3. "Under this view, today, banks want to hold roughly $700b in cash, and that has not changed despite the impovement in financial conditions."Banks can't change the aggregate quantity of reserve balances . . . only an offsetting change to the Fed's balance sheet can do that, by accounting identity. (The qualifier is when banks purchase vault cash with their reserve balances, since both are on the Fed's balance sheet.)If banks buy Treasuries, this just shifts the existing quantity from bank to bank (from the buying bank to the bank of the seller). Banks could buy at auction and drain reserve balances under current Treasury procedures (normally, though, the Treasury just re-credits the exact same amount into its tax and loan accounts to offset the reserve balance drain), but the Treasury's already sold all the securities it wants to, so even more banks buying at auction wouldn't have drained anymore balances than already has occurred.The Fed can control or otherwise target the qty of reserve balances (excess reserves are only reserve balances . . . vault cash doesn't count as excess reserves in the Fed's definition of the latter) if they set the target rate equal to the remuneration rate on reserve balances, which is what they currently have done. Japan did the same under so-called quantitative easing (which it wasn't), setting the target rate (zero) equal to the remuneration rate (zero). Otherwise, in order to hit its overnight target, the Fed has to drain or add balances (via standing facilities) to avoid the rate falling to the remuneration rate or rising to the standing facility rate, respectively.So, the existing quantity of reserve balances reflects Fed actions, not desired holding by banks.Best,Scott

  4. The quantity of reserve balances reflects Fed actions, but not the type.My point is that banks choose to hold cash (and by "cash" I mean deposits at the Fed) based on their liquidity preferences. They could easily choose to lower credit standards, and the result would be a shift from excess to required reserves. Banks in aggregate have about $17tr in on and off balance sheet assets. Can we be so sure that the roughly $700b in Fed deposits is "excessive" given the recent drop in the liquidity of those assets? Perhaps $700b, or roughly 4% of assets, in Fed deposits is the "new normal". Support for this view comes from the fact that, despite the dramatic reduction in credit spreads, banks have not lowered credit standards so as to lend out Fed deposits. One would have expected, on the margin, some change over the course of 2009 if Fed deposits were driven solely by a lack of lending opportunities. So my point is that if the Fed attempted to drain reserves, its not clear that banks would want to hold less on deposit at the Fed. Banks do not suffer a drain in "excess reserves" by Fed actions. Instead, they experience a drop in deposits. They can respond to a drop in deposits by reducing lending (reducing required reserves), or by reducing the amount on deposit at the Fed. In this recent piece, David Wheelock of the St. Louis Fed makes the point that the 1937 Fed should have drained Excess Reserves by selling assets rather than raising reserve requirements — he belives the impact on the money supply would have been much smaller. He makes an important caveat: "the impact would still have been large, however, if banks held excess reserves mainly as protection against depositor runs, rather than because they lacked profitable lending opportunities."http://research.stlouisfed.org/publications/mt/20090901/mtpub.pdf

  5. Absent remuneration at market rates, banks hold excess balances overnight for one reason only . . . as a buffer against needing to borrow from the central bank overnight at a penalty to clear an intraday overdraft. With remuneration at a market rate (that is, equal to the fed funds target), there is no opportunity cost to holding excess balances for an individual bank, since the return to lending to other banks is equal to the return to holding them in the Fed account. Therefore, there is no bidding of the fed funds rate down in the latter case when excess balances in the aggregate are greater than banks desire to hold for the above reason (avoiding borrowing at the penalty rate). But they still wouldn't necessarily desire to hold much more excess balances than without the remuneration, since their purpose hasn't changed and the balances necessary to settle the day's payments hasn't increased. To find out, the Fed could simply drain the excess balances until there was some pressure on the fed funds rate to rise or banks started turning to the Fed for overnight loans to avoid overdraft penalties, since this would indicate that the aggregate level of excess balances had fallen below banks' desired buffer level.It’s very clear from the international literature on bank reserve behavior that the decision to hold excess balances and lending standards are completely unrelated. Loans don’t necessarily end up creating reservable liabilities, anyway, as the liabilities created may ultimately be held as savings accounts, money market accounts, time deposits, corporate sweep accounts, and so forth.If you do something to change the relative cost of holding excess reserves compared to required reserves (i.e., the excess reserve tax), then, again, it’s highly doubtful this will alter lending standards. Instead, banks would most likely just attempt to re-label as many existing liabilities as possible as reservable. For instance, there are about $800 billion currently in retail sweep accounts currently classified as MMDAs (overnight, then deposits by day) to avoid reserve requirements . . . banks would likely just stop sweeping these balances and classify them as pure deposits. They would further incentivize customers to shift existing balances to reservable liabilities, given their own costs if customers do not do this. Given that we’ve assumed borrowers are no more likely to repay loans, this is highly preferable to lowering lending standards, raising required loan loss provisions as lending increases, risking greater capital loss (more loans but no better creditworthiness), and inevitably encouraging greater regulatory scrutiny.Best,Scott

  6. Scott,Thanks for the replies. I agree with much of what you say about lending standards and ER's. However, I also think you leave room in your reply for the same concept I am advancing: that bank "buffer stocks" drive bank liquidity decisions. What changed from 2008 on is that the perceived cost of having to borrow an excessive amount of funds in Fed overdraft facilities rose, or at least seemed to. Its not clear why this would be so — the Fed did not impose high non-monetary penalities on large overdrafts, and if anything, it went out of its way to remove the stigma of accessing them. However, the lessons of Bear Stearns, Lehman and Citi were clear: illiquidity was a grave risk to management independence and continuity. Surely this had an impact on the optimal amount of buffer needed to eliminate management career risk.So the point, I believe, stands: if the $700b in current ER's is seen as "optimal" or "necessary" by bank managements, then the impact of Fed withdrawal of TOTAL reserves is unclear. Take the case where the Fed sells $50b of MBS to a bank customer. In payment, the Fed debits the bank's reserve account. But what if the bank wants to maintain its reserve account at previous levels? Then it must reduce lending to produce the $50b and deposit it with the Fed (the same as it would if the Fed raised reserve requirements by that level). The impact of Fed asset sales on Excess Reserves, therefore, depends on the optimal buffer deposits banks want to hold at the Fed. The impact could therefore well be zero. The problem is also that "optimal level" could be highly volatile. In a situation where banks regain confidence, they may elect to lend out those reserves and keep less buffer, in part because assets would be seen as more liquid, and also because borrowers WOULD be more likely to repay loans. In this case, the reserves may be lent faster than the Fed can or is willing to raise the interest rate on excess reserves.I think the Fed is grossly overestimating the effectiveness of their future "Exit Plan" tools. For something so important, it seems to receive little critical review from the vast majority of monetary or macro economists. Your site is the exception in that it tends to question widely-held views on monetary topics.

  7. Hmmm…Reserves on deposit with Fed + Vault Cash = Excess + Required reserves = Total reserves If the Fed drains reserves, banks have no choice but to hold fewer reserve deposits at the Fed, unless they convert an equal amount of vault cash to Fed deposits.Also, the Fed can drain reserves without affecting the quantity of bank deposits by selling assets directly to banks.

  8. Anonymous,The Fed controls Total Reserves, not their composition. Banks control the composition, or rather, at any level of credit standards, their customers do.If the Fed drains reserves, banks need to hold fewer reserves. They can choose between "paying" the Fed by drawing down excess reserves, or by liquidating loans. The Fed does not control that choice — banks do. If the Fed sells assets directly to banks, again, they have a choice as to how to "pay": Draw down loans or use existing Fed deposits. In the case where they liquidate loans, deposits fall and so does the money supply.Look at it from the bank management's standpoint: its deposit at the Fed is just quasi-cash, not "excess reserves". The question is, how much cash does the bank management want to hold? The answer is independent of Fed actions, and dependent on perceived liquidity risk. This is not a smooth dynamic over time, but a "one-time" shift in liquidity preferences, and so it will not show up in any econometric analysis whose data set does not include material financial crises.Please find me a bank management that views its liquid asset position as Fed-determined. They don't say, "let me call up the Fed and ask them how much to deposit there today." Management can deposit any amount it wants at the Fed at any point in time merely by liquidating loans. If you assume loan levels are fixed in the presence of Excess Reserves, then you are absolutely right. This would be, in my mind, a very dangerous assumption.

  9. Hi David . . . getting closer and agreeing on more . . . appreciate the discussion. Had to split this one up due to the silly character requirement.“However, the lessons of Bear Stearns, Lehman and Citi were clear: illiquidity was a grave risk to management independence and continuity. Surely this had an impact on the optimal amount of buffer needed to eliminate management career risk.”It’s not illiquidity risk (the Fed guarantees any payments sent from reserve accounts whether or not a bank has a positive balance) as much as it is counterparty risk or concern with the smooth functioning of the payments system. The only thing ER do is provide that buffer, and the only reason for the buffer is because banks don’t have perfect certainty of clearing overdrafts by the end of the day . . . unlike Canada, where banks hold no ER overnight because they have perfect certainty of clearing overdrafts at the close of business. The payments system in the US is much more decentralized, though, creating a bit more uncertainty, and this plus the stiff penalties on overnight overdrafts and requirement of collateral for avoiding such an overdraft lead banks to hold a buffer here. But the size of the buffer has almost nothing to do with asset quantities or off-balance sheet commitments, and almost everything to do with the structure of the payments system (aside from the fact that banks with more assets will probably settle more payments, but even that doesn’t matter for ER in Canada).When 9/11 hit, the functioning of the payments system was affected, and for several days, banks desired to hold over $100 billion in ER as a result to ensure they could settle their payments. In fall 2008, banks obviously desired to hold more even than that, as the magnitude of the crisis was greater and spread to counterparty risk. Even then, though, the quantity they desired to hold was smaller than what the Fed provided (the Fed’s reasoning was that they didn’t have enough assets to sell to sterilize their substantial interventions) since the Fed funds rate traded well below the Fed’s target, often close to zero. Even when the Fed started paying interest on reserve balances, the funds rate traded near zero because some non-banks with reserve accounts were prohibited (until this spring) from earning interest, and the banking system didn’t want to absorb even their relatively small qty of balances.Now, with the payments system back essentially to normal functioning, the Fed could probably drain 90% to 99% of the ER without seeing any effect on the fed funds rate, since banks under normal circumstances desire to hold about $2 billion ER on average. The number could be a bit higher now given some continuing uncertainty about markets and also given interest payment. Banks were holding $2 billion ER 1 year ago under normal functioning of the payments system with no remuneration. It’s quite strange to suggest now, with mostly normal payments system functioning and the remuneration only 0.25% more that somehow they now want to hold $700 billion. That’s an exceptionally elastic demand for ER and just isn’t reasonable.Best,Scott

  10. Me again . . .“Take the case where the Fed sells $50b of MBS to a bank customer. In payment, the Fed debits the bank's reserve account. But what if the bank wants to maintain its reserve account at previous levels? Then it must reduce lending to produce the $50b and deposit it with the Fed (the same as it would if the Fed raised reserve requirements by that level).”That’s not what happens at all. The bank settles the $50b transaction, probably via an overdraft on its reserve account (at least partly), and then borrows the balances in money markets or from the Fed to cover the overdraft. If there’s no overdraft but the bank wants to hold more balances, they do the same thing—borrow them in money markets or from the Fed. If there aren’t enough balances to go around, the fed funds rate is bid up until the Fed accommodates to bring it back down (the Fed tries to anticipate this and carry out such an operation beforehand, actually). The bank’s decision to create loans or not has nothing to do with this sort of reserve management behavior. Also, if the Fed raised reserve requirements, the Fed would have to accommodate banks with more reserve balances if it is going to hit its fed funds target—so, again, there would be no change in bank lending related to reserve management behavior, aside from the reduced profitability to the bank of holding more reserve balances (but again, lending doesn’t necessarily mean more reservable liabilities).Best,Scott

  11. Last one . . . “Please find me a bank management that views its liquid asset position as Fed-determined. They don't say, "let me call up the Fed and ask them how much to deposit there today." Management can deposit any amount it wants at the Fed at any point in time merely by liquidating loans.”Just to be clear (this was a response to anonymous, not me, so may not have been intended for me), that’s not my point at all. My point is that—to agree with you, I think—changes to aggregate qty of reserve balances are set only by changes to the Fed’s balance sheet. Banks in the aggregate and individually definitely do set their own RR based upon their lending, to again (I think) agree with you. But the aggregate quantity of overnight ER after that is set by what the Fed does with total reserve balances. Under normal circumstances where the fed funds rate is set above the remuneration rate, the Fed accommodates banks’ preferences for both or else it can’t hit its target rate. With the remuneration rate set to the target rate, the Fed can supply any level in excess of banks’ desired combined levels of RR and overnight ER without seeing the fed funds rate bid below its target.Also, as a minor point (and maybe not relevant to your point), current regulations mean that RR are fixed 17 days before the beginning of any given two-week maintenance period. So, in real time at least, there would be no such thing as changing relative values of RR and ER within a maintenance period in the aggregate or for an individual bank.Best,Scott

  12. Scott,Thanks again for the replies. I think the difference in our views may come from assumptions about Fed overdraft facilities and other short term sources of financing for banks.If you assume a relatively friction-less use of Fed overdrafts and/or interbank loans, at a Fed Funds rate of zero, then there is never a need for banks to maintain liquidity in the form of Fed deposits. If the bank runs short of cash at end-of-day, as you say, it merely borrows from the Fed (or rather, the system does if it ends up short of funds in aggregate). The only purpose for excess Fed deposits would be for the type of operational clearing risk that you describe.My thesis, however, is that bank managements are wary of having to draw on Fed overdraft facilities or interbank loans in the event of another liquidity crisis. The reason is clear: drawing on Fed assistance jeopardizes management independence. Banks would rather sacrifice a small amount of lending profits to guarantee that they do not suffer the fate of Lehman. The (opportunity) cost of maintaining management independence is the lending spread minus reserve interest times reserve levels. For most banks, this is on the order of 15-20bps of ROA (including off-balance sheet assets), which, given the circumstances, might be seen as acceptable. So, to clearly state my assumptions:1) In the event of a Lehman-style liquidity crisis, use of Fed overdraft facilities (and the loans to clear them) carry significant risk to management independence.2) The inter-bank lending market is not an outlet for excess funds given credit risk.3) Banks want a 4% buffer of Fed deposits to support their business as long as 1) and 2) are trueOne last thing: the "buffer" level is a lower bound on bank Fed deposits, but obviously not an upper bound. That is, if the Fed wanted to grow reserves by another $1tr, this would end up in ER's. That does not mean, however, that withdrawal of $100b will come out of ER's. It may come out of lending.Keep in mind that historically, the Fed thought it prudent for banks to maintain much higher reserves. What happened was that views of market liquidity drastically improved, rendering reserve requirements archaic. Of course, we now know the view of what constitutes "normal" liquidity was, to put it mildly, misguided. The Fed has not changed its mind — reserve requirments haven't budged. Could it be, however, that banks themselves are re-definging what is "prudent"?

  13. Dear David . . . thanks to you, again . . . closer still (?)“If you assume a relatively friction-less use of Fed overdrafts and/or interbank loans, at a Fed Funds rate of zero, then there is never a need for banks to maintain liquidity in the form of Fed deposits. If the bank runs short of cash at end-of-day, as you say, it merely borrows from the Fed (or rather, the system does if it ends up short of funds in aggregate). The only purpose for excess Fed deposits would be for the type of operational clearing risk that you describe.”Don’t confuse liquidity (the ability to refinance) with settling payments, or, otherwise stated, be sure to distinguish between them. The reserve balance “buffer” is under normal circumstances for the uncertainty related to the latter, not the former. Liquidity risk has to do with being able to refinance short-term liabilities without rates rising considerably or having to sell off assets into a declining market. This (lack of liquidity risk) is normally ensured by close arbitrage of other money market rates with the Fed’s target.The difference between 9/11 and the recent crisis is that the former was mostly about payments and the latter was about mostly liquidity (as spreads from the Fed’s target went through the roof), but eventually both when some of the non-bank payments/collateral infrastructure went down with Lehman. What banks need is assurance that they can refinance their liability side without rates rising . . . that was the point of the various “credit easing” standing facilities set up by the Fed. Now the Fed has wound down most of those facilities, while the payments system has been back to normal for quite some time.Certainly either of these risks could raise banks’ desired “buffer,” but, again, it was clear last fall that banks didn’t even want to hold $600 billion at the height of the crisis as the fed funds rate fell to essentially zero many times. Now, without either risk being substantial, it’s hard to fathom they would desire to hold still more than that. This would be my primary disagreement with your three assumptions.“My thesis, however, is that bank managements are wary of having to draw on Fed overdraft facilities or interbank loans in the event of another liquidity crisis. The reason is clear: drawing on Fed assistance jeopardizes management independence. Banks would rather sacrifice a small amount of lending profits to guarantee that they do not suffer the fate of Lehman.”OK. Mostly agree, as I noted above. “That does not mean, however, that withdrawal of $100b will come out of ER's. It may come out of lending.”Unclear what is meant here.“Of course, we now know the view of what constitutes "normal" liquidity was, to put it mildly, misguided."I think the problem is that the Fed doesn’t understand its own purpose or its own monetary operations. The fundamental role of a central bank is to minimize payments and liquidity risks to the financial system; that’s why they were originally founded. If they do this, it would leave a minimal buffer, if any at all, as in Canada. If the Fed did two things, it could avoid both risks, even in a crisis, at least to the banking system (and others, depending upon whether one prefers to extend these proposals to non-bank FIs—pretty controversial to wade into those waters, though). First, eliminate uncertainty regarding overnight overdrafts . . many, many ways to do this, and Canada’s method is just one example. Second, create standing facilities for banks out to six months maturity or more (the Fed tried to do this with the TAF, CPFF, cb swaps, treasury lending facility, etc., but never got it quite right and didn’t move fast enough); no need to “discipline” banks on the liability side since their deposits are government insured anyway . . . regulate the asset side instead. Best,Scott

  14. Scott,Thanks — this is difficult issue for the layman to understand, and your replies cast considerable light on it. Just one other set of questions: do you believe that withdrawal of excess reserves would have any impact on aggregate demand? If not, why all this talk about a "difficult exit" from the Fed's balance sheet expansion? The Fed says don't worry about inflation because it will withdraw excess reserves before they are lent out. But if the excess reserves haven't been lent out (by definition), then how could withdrawing them have any impact on demand? Why not just withdraw them tomorrow? Obviously, my thesis on liquidity buffers was an attempt to solve the above conundrum. In the absence of that explanation, I'm left confused as to 1) why the Fed created excess reserves if not to provide banks with a liquidity buffer; and 2) why it doesn't just withdraw them tomorrow? Best,David

  15. DP:My view is that the Fed has created excess reserves primarily because of their usefulness as a Fed balance sheet item, as opposed to their usefulness as a commercial bank balance sheet item.Leaving aside the idea that loans create deposits for a moment, the usefulness of reserves to the Fed has to do with their role as a source of funding or a balancing item on the right hand side of the Fed balance sheet. This allows the Fed to expand its balance sheet directly through its interconnection with the commercial banking system. The Fed has had an interest in expanding its balance sheet in the context of its “credit easing” program, which involves acquisition of private sector assets. Asset acquisition creates new reserves.(This central bank dynamic very much parallels that of the creation of new reserves at the outset through government spending, which is the one more typically discussed in this blog.)One alternative to using excess reserves in this way is the issuance of treasury bills by the government, who would park the resulting funds in its account with the Fed. This would drain the excess reserves and replace them with a government deposit. They did this for a while (in fact $ 200 billion still remains in what is called the “US Treasury supplementary financing account”), but it becomes politically sensitive for the Fed to be using the government as its agent in this sense. It’s a hot point for the perception of the “independence issue”, and it’s even a role reversal in terms of the normal agency relationship. A second alternative is that the Fed could also issue its own interest paying liabilities, but they haven’t done so yet. (I don’t know whether they have the authority to do so yet. If they don’t, they’ll probably get it soon.)The reason the Fed would not withdraw reserves tomorrow is that it’s not ready to shrink its balance sheet back to normal size yet, and that’s because it’s not ready to wind down the various credit easing programs yet. In fact, some people think they could still be going in the other direction – additional asset expansion – if the housing market remains stalled, and they want to acquire more assets. Because it’s not ready to unwind, it requires a sustained source of “funding” in addition to outstanding currency.Excess reserves were useful and necessary for the commercial banking system, to a degree, particularly in the earlier stages of the crisis. But there’s no way the commercial banking system is “demanding” $ 800 billion in excess reserves simply due to liquidity oriented caution, even in this environment.The fact that the Fed is using reserves mostly for their own purpose is also evident in the fact they are paying interest on reserves. This acts as a disincentive for commercial banks to be tempted by reckless, dark side, multiplier thinking (i.e. wrongheaded), in terms of how to respond to the new reserves created by the Fed, reserves that the banks don’t really need, and that are more for the Fed's own purpose. If the banks are receiving interest on reserves they’ve collectively had no control in creating, they will be more inclined to continue to implement their credit policies base on proper risk assessment and appropriate capital adequacy levels, as they should do.

  16. JKH,I understand your point, but I have a counterargument:How does "credit easing" help the banking system if the funds injected are not used? Let's say the banking system had "too many" MBS post-Lehman. "Too many" means, in my mind, that they could not liquidate MBS at an acceptable price. So they went to the Fed and liquidated them in return for Excess Reserves. However, Scott says the banks don't need Excess Reserves. So why would the banks liquidate MBS in exchange for something they don't need? And why would the Fed want to expand its balance sheet to give banks something they don't want?Answer: because the banks did, actually, need a liquidity cushion. The purpose of "credit easing" was to provide a cushion that banks would have created themselves, by dumping MBS into already-fragile markets. The problem is that Scott believes the banks don't need Fed deposits as liquidity buffers — they can just run overdrafts at the Fed when they system comes up short of cash, and they can cover these overdrafts with Fed loans. So, if I understand Scott correctly, there was no need, or at least no need of any permanence, for the Fed to provide the banking system with excess reserves. This is the essence of my disagreement with Scott. If the banks did need a liquidity buffer, and the Fed did engage in Credit Easing to provide one, then one can reasonably conclude that the banks desire to have excess reserves at these levels, and they will seek to maintain those reserves at the expense of required reserves — i.e., at the expense of lending.So if banks don't need the liquidity buffer, the excess reserves serve no purpose — the Fed can sell back the MBS in exchange for unneeded Excess Reserves. And if the banks do need the buffer, and the Fed tries to shrink its balance sheet, then the money supply would tank.

  17. David,First, there’s a difference in the banking system effect, depending on whether the Fed acquires assets from a bank with a reserve account at the Fed, or from a non bank without a reserve account. The first type of transaction merely replaces a bank asset with additional reserves; the second type actually expands the size of the banking system by creating new reserves and new deposit liabilities. And banks may also be acting in a broker function for the second type, where they have the asset on their own books for only a short period of time. The Fed has numerous asset acquisition programs going on and numerous counterparties to deal with. I haven’t looked closely enough at the whole thing to attempt to guess on the split among these various types of transactions and the net effect on the overall banking system profile relative to the counterfactual of no Fed intervention. That would seem like a near impossible task in any event. That said, let’s assume as in your example it is a bank that sells the asset to the Fed (or effectively acquires financing for the asset from the Fed).Second, my interpretation of credit easing is that the Fed made a judgment that the credit markets just weren’t functioning in terms of reasonable pricing and liquidity of execution. That’s an extreme understatement if anything. Banks and others couldn’t sell or place these various assets at reasonable terms – the market basically collapsed due to extreme liquidity and credit risk paranoia and general pricing chaos. It is important to note though that if the system had been able to function without Fed intervention, there would have been no effect on excess reserves. The Fed determines the level of excess reserves.Third, the origin of the problem was not that the banking system was seeking excess reserves. The system in its entirety was just looking to sell or finance these various assets under conditions that were more normal than what was obviously the case. This mostly has to do with pricing, in effect. In normal circumstances, if pricing can be achieved at reasonable terms, there is no problem with the availability of excess reserves for the system, by definition. Transactions get completed, reserves get redistributed as necessary, and everybody’s happy. The problem is when transactions can’t get completed according to wishes and everybody isn’t happy.- continues next

  18. – continued from previousThe Fed intervened in order to alleviate this logjam. It provided cash through its various credit easing programs. That increased excess reserves as a necessary consequence of its intervention. The Fed couldn’t intervene as it did without eventually expanding its balance sheet, and it needed some form of additional liability expansion to do so – it choose to do this in the form of excess reserves.Return to your example, where a bank sells an MBS to the Fed. Why did it sell? To get excess reserves? No. It sold because the Fed was making a market, meaning making a price at which it was willing to transact, and a price that would not have been available to the bank without the Fed’s intervention in the market.In fact, this bank might not end up with excess reserves at all, as a result of that transaction. It depends on the counterfactual. Suppose some greedy pension fund had already been providing some form of money market financing to the bank for that position, at some extreme market rate. The Fed judges that sort of predatory pricing to be part of the market dysfunction, and offers a better deal. The Fed pays the bank and the bank collapses its financing from the greedy pension fund. The bank ends up with no excess reserves.But somebody does. The money retraces some route back through the pension fund and some bank must end up with excess reserves.Was that bank looking for excess reserves? No.Similarly, the entire system ends up with excess reserves that it wasn’t really looking for. And the reason is that excess reserves are a necessary by-product of Fed intervention. The Fed’s objective was to break the market dysfunction in pricing and risk transfer; not to provide excess reserves per se – at least not to the level of $ 800 billion. $ 800 billion in excess reserves is the necessary consequence of its asset intervention, not the reason for it.In summary, credit easing provided much more than “the cushion that banks would have created themselves by dumping MBS into already-fragile markets”. The banks could not have created excess reserves collectively by doing that. Only the Fed can do that. But it’s the pricing and transaction flow effect that could only have been achieved by Fed intervention. And it was necessary to create excess reserves in order to achieve the level of desired intervention, and the desired pricing influence. But the banks don’t need those reserves.As far as the overdraft issue is concerned, there are several differences between that back door approach and the front door approach that the Fed designed more formally through credit easing. First, the Fed structured a number of different programs with different credit qualification objectives that wouldn’t necessarily be feasible through normal window facilities. But more generally, even if the collateral arrangements had been feasible through the discount window, the general chaos in the market required a much more structured and front door (non-stigmatized) approach.

  19. Dear DavidI agree with JKH's description. He said virtually everything I would have said, perhaps better. The only thing I would add is that the Fed left excess reserves circulating beginning after Lehman because it didn't think it had the assets to sell to drain them (JKH points out some alternatives to selling securities, though some of these aren't authorized by Congress, unfortunately in my view). This (large excess reserves) wasn't the case prior to Lehman, as the "credit easing" actions starting in December 2007 through Lehman were completely sterilized by sales of Treasuries (mostly). In short, then, the Fed believed that the reserve effects of the larger credit easing actions necessary after Lehman could not be sterilized like those prior to Lehman were. It's also no mistake that the Fed started paying interest right after Lehman, since they believed they couldn't drain the excess reserves but still wanted to achieve a positive target rate. They bumbled that one, though, only getting it right this spring. Luckily, it didn't really matter since they set the target to nearly zero in short order. As far as the point of MBS purchases, and indeed all of the "credit easing" operations the only point is to set a price (yield) in these markets lower than what is set without the Fed's intervention. That is, the only effect of the Fed's purchases of MBS, Treasuries, and any other asset is to raise the price (and obviously lower the yield). That's what the Fed's after . . . reduce interest rates in these markets since obviously having the fed funds target at zero isn't bringing these rates down anymore. (As an aside, the Fed doesn't understand that they could be much more effective at this by simply announcing the yield they are after or the desired spread above the benchmark Tsy rate; consequently, these purchases haven't been as effective as they otherwise might have, in my view.) The reason the Fed doesn't sell these MBS and securities to drain the excess reserves is because it would reverse the price effects of buying them. But they do believe they will do this once the markets are back to "normal" functioning, which is generally defined as returning to normal spreads (which would indicate market activity back to normal . . I've argued that it doesn't matter if they do not, at least in terms of whether they leave the excess reserves circulating or not).The other credit easing actions are similar even though they are loans rather than securities purchases. For instance, the TAF and CPFF simply set rates for refinancing short-term liabilities, since these are the markets that weren't functioning properly (again, spreads way above normal). Here again, the Fed's mistake (at least for the TAF, they got it more right with the CPFF) was in not setting a rate, but allowing that to be set via auction while limiting avaiable loans.Best,Scott

  20. JKH's two-part reply wasn't posted yet when I wrote mine. I agree completely with it.Best,Scott

  21. Scott and JKH,Lacker seems to agree with me, as today he stated that there may be no need for completing the MBS purchases as "bank reserve demand (is) ebbing as financial conditions improve."I think we're arguing past each other a bit. The Fed IS trying to bring mortgage spreads down. Why were those spreads wide? Because of the perception that banks would dump mortgages in an effort to raise liquid funds. The Fed took away that perception by provinding those liquid funds in exchange for mortgages. I'm not sure I fully understand JKH's point on the impact of Fed MBS sales on mortgage spreads. If one assumes that there is zero cost to banks being drained of excess reserves, then they would gladly be willing to trade those reserves for mortgages at current prices. Imagine the discussion between the NY Fed and the Bank CFO.NY Fed: I want to swap $50b mortgages for $50b in resrerves at yesterday's close.Bank CFO: Done. I didn't want those reserves anyway.When you say that "mortgage spreads would rise", you imply a there is a value to banks holding reserves instead of mortgages. What is that value? Where does it come from? Buffer liquidity, of course. Either reserves have value to banks or they don't. If they don't the banks will gladly buy mortgages from the Fed at any price (yield). If they do, then the value comes purely from the buffer liquidity they offer.BTW, you may argue the banks don't want MBS credit risk. Let's leave aside the very-real existence of a Treasury guarantee. If the banks don't want that risk, they can sell mortgages and buy Treasuries. This would have the effect of widening, somewhat, the mortgage spread, and reducing, somewhat, the Treasury yield. The overall impact on mortgage rates? Difficult to estimate, but certainly not a wipe-out for the housing sector.

  22. BTW, the credit risk issue disappears if the Fed sells Treasury bonds instead of MBS. If you assume that banks don't want longer-maturity assets, then you are making an implicit assumption about bank liquidity preferences. Take that assumption to its logical conclusion:1) banks don't want Treasury Notes or Bonds because they are less liquid.2) banks want to hold short-maturity assets.3) post-lehman, shorter maturity assets carry too much liquidity risk, otherwise the banks would have already bought them with their ER's.4) banks would rather hold excess reserves than any other short-maturity asset.

  23. It appears I used Bloomberg's paraphrasing of Lacker, but what he actually said in his speech supports my point:"In response to heightened creditor anxiety, many banks build up large buffers of highly liquid assets…some of which they held in reserve balances at the Federal Reserve Banks."and, more specific to my argument:"…those [the Fed's MBS] purchases supply reserves which reduce the amount the banks need to borrow from the Fed to satisfy their elevated demand for reserve account balances."Lacker is arguing that ER's constitute the liquidity buffers of banks. The implication is that draining those reserves would cause banks to borrow from the Fed to restore them, forcing the Fed's balance sheet to re-expand. To prevent that from occurring, the Fed would have to raise the Fed Funds rate, perhaps dramatically. Thus, in Lacker's view, presumably there is no "exit" from ER's held as liquidity buffers without a corresponding increase in the Fed Funds rate. Of course, this would be highly contractionary. So we get back to the orginal thesis: as long as banks want to hold ER's as liquidity buffers, the Fed cannot drain them without a significant negative impact on the money supply.

  24. David,You’re right. Lacker devotes an unusually sizeable chunk of his speech talking about demand for excess reserves:http://www.richmondfed.org/press_room/speeches/president_jeff_lacker/2009/lacker_speech_20090827.cfmHe’s on the FOMC, so he should know. But I disagree almost entirely with his interpretation. I say almost because I think the reserve demand argument was much closer to the full truth at the very outset of the crisis, when the market was in state of complete collapse, post Lehman, and banks were fearful of lending to each other. This became less the case as the Fed rolled out various programs and the markets gradually regained their footings.The fact is that the latest Fed balance sheet shows $ 800 billion in excess reserves and only $ 30 billion in primary credit (essentially borrowed reserves). It’s difficult to make the argument that $ 800 billion is necessary when window borrowing is only $ 30 billion. That seems to be the argument he’s making, which I think is nonsense. He further seems to say that excess demand for reserves will finally be satisfied when window borrowing has ceased. I also think that’s nonsense. I think window borrowing reflects the demands of specific problem institutions. That’s different from systemic excess demand. The explanation in my view is that discount window borrowing reflects specific bank cases within the still shaky state of the US banking system from a solvency perspective, particularly with respect to the regional banks. It is impossible to completely disengage the issue of “liquidity risk” as per discount window demand from solvency risk as per market perception of particular institutions.That, plus the fact that the Fed is on record as saying that these various programs are designed to target particular problem credit areas. They didn’t mention reserve demand at all when they launched these programs, as far as I can recall.Remember that these Fed members when they speak generally do so on their own behalf, and say so in the introduction. They don’t speak on behalf of Fed policy. Only Fed releases do that. So Lacker’s interpretation of what is going on isn’t necessarily held by all Fed members (although it may be), and nor is it necessarily correct.In fact, he says in his introduction:“Please note that these are my own views and should not be attributed to any other person in the Federal Reserve System.”Bottom line is I disagree with his interpretation.But what he says certainly supports your view.David Pearson, are you in truth Jeffrey Lacker?

  25. David,On the other point, the Fed will start winding down its asset programs when it’s ready:My version of the conversion:Goldman Hot Shot:Hey Fedster, we’re bullish on mortgages over here. We’re prepared to bid you for your entire book.Fed Man: Sorry, Goldie. We’re not offering yet. Still some problem areas out there. Why not give Citi a call? They may have an interest. And I suggest you do the same with the FDIC. They’ve got some to go after last weekend’s roundup.Goldman Hot Shot:But we can help you clean up your excess reserves.Fed Man:We’re not ready to clean them up.Aw, Fedster. You’re no fun.Fed Man:Make my day, punk.

  26. Correction: Term auction credit on the Fed’s balance sheet is another $ 221 billion.So the effective total borrowed is about $ 250 billion.But if those who have borrowed ended up with flat reserve positions, those who haven’t ended up sharing the $ 800 billion excess.In an alternative hypothetical scenario, the Fed might have targeted the $ 250 billion to those who needed it, and drained any excess that existed among those who didn’t.But it didn’t do that because it needed the $ 800 billion to offset its other asset programs.