Why (and when) interest-on-reserves matters…

Paul Krugman writes:

Incidentally, small nerdy note. Some people argue that the concept of the monetary base has lost its relevance now that the Fed pays (trivial) interest on reserves. I disagree. Reserves and currency are fungible: banks can turn one into the other at will. But the total of reserves and currency is fixed by the Fed — nobody else can create either. That, as I see it, makes them a relevant aggregate — and anyone who believes that all those reserves are sitting idle because of that 25 basis point reward is (a) silly (b) ignorant of Japan’s experience, where the BOJ sharply increased the monetary base without paying interest on reserves, and what happened looked exactly like our own later experience.

Nerdy indeed! Some might even describe it as dork-ish.

But I think that Krugman is mischaracterizing the view he is arguing with. I’m not sure he would even argue with the view properly characterized. Of course he might!

Perhaps there are people in the world who think that paying 25 basis points of interest on reserves means that base money doesn’t matter, but I have not met any of them. I certainly agree with Krugman that those 25 basis points have a pretty negligible macroeconomic effect now.

The view of people who think that interest-on-reserves permanently diminishes the macroeconomic meaning of base money is contingent on a conjecture that, henceforth, the Fed will always pay interest-on-reserves at a rate comparable to the rate paid by short-term US Treasury securities. If that conjecture is false, then the quantity base money will someday matter again.

In either case, the quantity of base money that exists right now is largely meaningless , and would be meaningless if the Fed were not paying any interest on reserves, because Treasury rates are near zero. This is Krugman’s liquidity trap, an effect of the negative unnatural rate of interest.

The quantity of base money is meaningless right now even if I am wrong about the Fed henceforth always paying interest on reserves at the short-term Treasury rate. Because if I am wrong, the only way that the Fed can create a spread between the interest rate paid on base money (whether zero or something higher) and the Treasury rate is to dramatically reduce the quantity of base money, so that some convenience yield on holding scarce base money offsets the opportunity cost implied by a T-bill / base money spread. Current quantities of base money simply can’t coexist with with a spread between the interest paid on reserves and the interest paid on Treasury bills. (Unless some unexpected thing dramatically increases the convenience yield of holding base money!).

So, there is almost no direct informational value to the current quantity of base money. Perhaps there is indirect information that matters. It could well be that the rate of change of the quantity of base money contains information about the likelihood of future interest rate changes, so it is not irrational of market participants to respond to rumors of tapering or moar QE. Perhaps there are institutional quirks related to the fact market participants can only hold interest-paying base money indirectly via banks, while the stock of risk-free securities depleted by monetary expansion can be held (and hypothecated) by non-bank actors. (If so, “monetary expansion” might be contractionary!)

But the first-order effect of monetary policy is gone. Changes in the base used to engender straightforward imbalances between a direct opportunity cost and the convenience yield of holding money. A reduction of interest rates / expansion of the monetary base would lead to an increase in the direct cost of holding money rather than Treasuries, and put the economy in disequilibrium until NGDP or (too frequently) asset prices adjusted to increase the convenience yield attached to the monetary base. A contraction did the reverse. While we are stuck at zero we can argue over expectations or collateral chains, but the old, blunt, simple channel no longer functions.

And it will never function, as long as the Fed always pays interest comparable to Treasuries on base money. There is nothing special about zero, or 25 bps. What makes a liquidity trap is that the rate of interest paid on money is greater than or comparable to the rate of interest paid on Treasury bonds. So long as that is true, whatever the level of interest rates of interest, macroeconomic outcomes will be much less sensitive to changes in the quantity of money than in once-ordinary times. That is not to say, full-stop, that monetary policy is impotent. Those squishy expectations and institutional quirks may matter. But post-2008, we live in a world where insufficiently expansionary monetary policy has meant tripling the monetary base. Pre-2008, tiny changes in the quantity of base were sufficient to halt an expansion or risk an inflation. The relative impotence of changes in the monetary base is not a function of the zero-lower-bound. It is a function of the spread between base money and risk-free debt, a spread which may well be gone forever.

 
 

63 Responses to “Why (and when) interest-on-reserves matters…”

  1. Detroit Dan writes:

    The effect of the Fed’s interest rate tinkering has always been short-lived and insignificant in the larger scope of things. Thus, we experienced 30 years of falling interest rates strongly correlated with falling inflation. Lower inflation rates caused by the Fed creating more base money did not cause inflation, other than for very short spurts which were soon reversed. Other factors (technology, globalization, “ownership society” policies) overwhelmed the trivial impact of monetary policy…

  2. Detroit Dan writes:

    I meant “Lower interest rates caused by the Fed creating more base money did not cause inflation…”. Sorry for the mistake…

  3. Detroit Dan writes:

    And of course the Fed always follows inflation. Monetary policy is insignificant all the time, except for the financial speculators, or when the Fed does something stupid like try to control the base money supply (bank reserves) as with Volcker around 1980 — see John Carney

  4. stone writes:

    JKH says that most of the treasuries bought by QE were not held by banks but rather held by non-bank institutions that then asked banks to sell them on their behalf to the Fed. So the people who were previously holding the treasuries end up holding bank deposits rather than directly holding bank reserves. Those people left with bank deposits have a problem because unlike treasuries, bank deposits are at risk (if over the insured limit) if their bank were to collapse. So I suppose that risky nature of bank deposits gives a bit of a “hot potato” effect from QE?

  5. […] * And never miss a new Steve Randy Waldman. […]

  6. Handle writes:

    As I understand it, a large portion for the explanation of the explosion in reserves and the tiny but above-zero interest rate is to ensure that Money Markets don’t break the buck, something that is thought to be prone to setting off a large-scale panic. A panic of … using cash sitting in Money Markets to bid up other investment assets?

    Money Markets are now at $2.7 Trillion. (Institutional: 1,750, Retail: 660, Other: 275) and pay less than inflation expectations. Trends in Retail and Industrial both look very correlated to housing-bubble prices and have stabilized.

  7. … Senaste Blogroll Tillsatser

    [ ... ] Wow , superb weblog struktur ! Hur långa har du någonsin varit igång en blogg ? För [ ... ]

  8. […] 2013 by Mark Thoma The Great Recession crushed America’s economic potential – Neil Irwin Why (and when) interest-on-reserves matters… – interfluidity Monetary Policy and Emerging […]

  9. Nick Rowe writes:

    Steve: I’ve read this twice, but I’m afraid I can’t understand exactly what you are saying.

    One point of information: for the last 20 years the Bank of Canada has been paying interest on reserves, and has had no reserve requirements (commercial banks can hold either positive or negative reserves if they wish, but mostly choose to hold very close to zero reserves, aside from small amounts of currency, mostly in ATMs).

    For example, right now the Bank of Canada: pays 0.75% interest on reserves (positive balances at the BoC); charges 1.25% on negative balances at the Bank of Canada: and has a target for the overnight rate of 1.00%. The actual overnight rate cannot fall below 0.75% (otherwise it would be more profitable to lend to the Bank of Canada) and cannot rise above 1.25% (otherwise it would be cheaper to lend to the Bank of Canada). But the Bank of Canada keeps the actual overnight rate very close to the 1.00% target by threatening to add or subtract settlement balances if the actual overnight rate rises above or falls below the target.

    And every 6 weeks the BoC adjusts all three interest rates, as needed, but always keeps the same 25 basis point spreads.

    This policy means that the interest rate on reserves is always very close to the interest rate on short term Treasury Bills. But this does not mean that monetary policy is powerless. The Bank of Canada has been very successful in keeping inflation close to the 2% target over the last 20 years (despite some massive changes in fiscal policy that were done for entirely unrelated reasons and were offset by monetary policy).

    So when you say “But the first-order effect of monetary policy is gone.” I have no idea what you could mean by that. Because the effects of monetary policy have been very much present and correct in Canada for the last 20 years, despite paying interest rates on reserves that track the TBill rate fairly closely.

    And the *only* thing the Bank of Canada really controls is its own balance sheet (plus it controls the signals it makes about how it will adjust that balance sheet in future).

  10. Nick Rowe writes:

    Damn typo! That should read:

    “and cannot rise above 1.25% (otherwise it would be cheaper to BORROW FROM the Bank of Canada).”

  11. Dan Kervick writes:

    Stone wrote:

    “Those people left with bank deposits have a problem because unlike treasuries, bank deposits are at risk (if over the insured limit) if their bank were to collapse.”

    Question: So then why did those people sell their treasuries?

  12. Dan Kervick writes:

    SRW wrote:

    Because if I am wrong, the only way that the Fed can create a spread between the interest rate paid on base money (whether zero or something higher) and the Treasury rate is to dramatically reduce the quantity of base money, so that some convenience yield on holding scarce base money offsets the opportunity cost implied by a T-bill / base money spread.

    The Fed can’t control rates and quantity at the same time, can it? If the Fed reverses QE and sells off a bunch of securities to dramatically reduce the quantity of base money, thus increasing the demand for base money and driving up the overnight rate, the banks would borrow the money right back from the Fed unless the Fed allows the discount rate to float upward along with the demand. There can’t be an interest rate regime and a quantity regime at the same time.

  13. stone writes:

    Dan Kervick @11 “Question: So then why did those people sell their treasuries?”
    I suppose they sold because they thought that QE offered an exceptionally good price for the treasuries and also (a bit of a circular argument) other prices were surging and they wanted to jump onto those gravy trains????

  14. stone writes:

    Nick Rowe@9, in Canada, do you think that monetary policy has its effects purely through influencing interest rates?
    My impression is that Scott Sumner says that increasing the monetary base will increase NGDP even when the monetary base has already been increased so much that further increases no longer impact on interest rates and that that position is what this post is grappling with (sorry if I’m in a muddle or stating the obvious).

  15. Dan Kervick writes:

    stone @13, that sounds right to me. So I wonder why people who willingly make a purchase because they like the price would then decide that the thing they willingly bought is a problem for them, and a hot potato.

  16. […] Here’s Steve Waldman: […]

  17. JKH writes:

    In general, the topic suffers from a lack of clear differentiation in the economics of the two components of the monetary base. That is why the concept of the “monetary base” as a lump is toxic for efficient and effective analysis. And the mere fact that currency is exchangeable for bank reserves is almost irrelevant in analyzing this. The additional fact that some economic schools seem to do everything possible to deemphasize interest rates as an aspect of monetary policy doesn’t help.

    The convenience yield approach seems like an awfully indirect method of analyzing this.

    OT – has anybody noticed that Krugman has undertaken a somewhat stealth transition to Romer’s ISMP model over the past couple of months?

  18. Nick Rowe writes:

    stone: “Nick Rowe@9, in Canada, do you think that monetary policy has its effects purely through influencing interest rates?”

    Most economists see it that way. I don’t. I see interest rates as a symptom of monetary policy. But this argument is a bit like arguing whether that drawing is a drawing of a duck or a rabbit. It isn’t really one or the other. It’s more a question of what interpretation is more useful.

  19. stone writes:

    Nick Rowe@18, Do you doubt it is possible to decouple monetary base from interest rates such that the monetary base gets expanded (perhaps a lot) and yet all interest rates remain the same?

  20. JP Koning writes:

    Good post. My feeling is that increases in the base still have some effect, albeit an ever diminishing one, because they signal that the Fed is extending its expectations that the spread between IOR and the overnight rate, or the convenience yield, stay at 0 ever farther into the future.

    Nick: “So when you say “But the first-order effect of monetary policy is gone.” I have no idea what you could mean by that. Because the effects of monetary policy have been very much present and correct in Canada for the last 20 years, despite paying interest rates on reserves that track the TBill rate fairly closely.”

    It seems to me that if the BoC created enough settlement balances to drive the overnight rate down to the deposit rate (sort of like what it did in 2009-2010), and if this 0% spread was expected to be permanent, then subsequent open market purchases would do very little, as Steve points out. This isn’t to say that base money is irrelevant. The Bank can still exercise monetary policy by moving the deposit rate on base money up or down rather than adding/subtracting small amounts of balances so as to manipulate their convenience yield.

  21. Nick Rowe writes:

    stone: that is exactly what the Quantity Theory says will happen. If there’s a permanent doubling of the money base, all prices will double, and all interest rates will stay the same. It is also possible to get the same results, but with prices not changing either, if the monetary base increases temporarily but is expected to fall back down again a lot in the near future. And those possibilities are great examples of why it is not useful to think of monetary policy as interest rate policy.

  22. Nick Rowe writes:

    JP Koning: But that is like saying: “what happens if the central bank lends out a million dollars at 5% interest, and at the same time borrows a million dollars at 5% interest?” Nothing will happen (unless people see it as a signal of something). It’s just an open market purchase, coupled with an open market sale. The monetary base doesn’t expand.

  23. jt26 writes:

    JP/NR: re:monetary base and expectations. Riffing on JP’s blog posts, would an “empirical proof” be a plot of MB vs. Fed Funds futures?

  24. ezra abrams writes:

    I don’t undestand a word you said.
    However, the original argument seems pretty simple: banks are not lending cause they would rather have a safe 0.25% at the treasury.
    This means there are no projects that banks are willing to fund by loaning money at a rate R=0.25%+S, where S is the premium required to compensat the bank for risk over treasury.
    I don’t know what S is today; lets say 5%
    Are you really going to tell me that there are lots and lots and lots of loans banks won’t make at 5.25% ? and they would make them if the treasury paid zero ?
    Or, if you are a banker, I have 100,000,000.00 dollars at 0.25% minus inflation (250,000.00 per year) or I could loan out 50,000,000.00 at 5.25% (2,630,000.00 per year)
    I mean, they are really saying they can’t find or are unwilling to make loans and pocket an extra 2.5 mill a year ???????
    IF that is true, the problem is not the 0.25%; it is either the bankers, who are not doing thier financial intermdiation job, or the eocnomy, which is not providing demand….since we know tht corporations are sitting on huge amounts of cash, and that wage demands are highly moderated, the slack demand hypothesis seems reasoanble..
    isn’t that the CORE of the argument ???

    am I missing something ??

  25. stone writes:

    Nick Rowe@21, So in your view is the previous lack of inflation in Japan at least in part due to a belief that Japanese QE will be reversed (or taxes will be higher than government spending) such that the increase in Japanese monetary base gets reversed? What proportion of the population need to have such a belief for prices to get reset or not? My impression in the UK is that very few people have a clue what monetary base is or whether it has increased let alone whether that increase is permanent.

    I am struck though by how prices did not drift up during the gold standard times in C19th UK in the way that they have since. I suppose there were periodic depressions with debt write downs and that reversed prior price rises. Is the mechanism by which the monetary base resets prices through being a backstop whenever there are cascades of bankruptcies? If so then to my mind the relevant measure would be the sum total of the monetary base AND treasuries. Basically all risk free financial assets. BUT now the government seems in favor of full on bailouts whenever a crisis looms so perhaps there is an unlimited backstop. The backstop will be whatever it needs to be in order to protect wealth however that wealth was concocted.

  26. Mark A. Sadowski writes:

    To enlarge on the point Nick Rowe made with respect to Canada, that is that Canada has had interest on reserves since 1992, interest on reserves is currently in practice in several currency areas that comprise over half of global GDP.

    In particular, to the best of my knowledge, Norway has had it since 1994, Sweden since 1994, Australia since 1997, the Euro Area since 1998, New Zealand since 1999, the UK since 2001 and Japan since 2008. With respect to explicit inflation rate targeting, New Zealand has practiced it since 1990, Canada since 1991, the UK since 1992, Sweden since 1993, Australia since 1994 and Norway since 2001.

    Of these currency areas only the US, Japan and the UK pay an interest rate on reserves that exceeds the lowest rate on government bonds. In the rest of these currency areas the rate on reserves is typically on a little below the lowest rate on bonds. These three currency areas also happen to be the only ones currently clearly at the zero lower bound in policy interest rates, and which also currently have QE programs. I don’t think anyone would claim that any of the other currency areas are in a liquidity trap with the possible exception of the Euro Area, but I think a good argument can be made that the Euro Area’s monetary policy problems are extremely self inflicted.

    And although Canada has a record of unsually low inflation variability, I don’t think any of the currency areas explicitly practicing inflation rate targeting have had much difficulty in hitting their targets, and that even includes the UK which has been above target for much of the time it has been at the zero lower bound in policy interest rates.

    And as a final note, with the exception of the US, the Euro Area, Japan and the UK, the monetary base in each of these currency areas is currently only about 3-4% of GDP, which is actually pretty low by historical international standards.

  27. Mark A. Sadowski writes:

    Minor correction:

    Norway has had interest on reserves since 1993.

  28. Nick Rowe writes:

    Here’s the general way I think about all this:

    Let’s ignore the distinction between currency and other reserves, and just assume the central bank issues currency and pays interest on that currency, and can vary that rate of interest whenever it chooses. That gives the central bank two instruments: it can vary the supply of currency, and it can vary the demand for currency by changing the rate of interest it pays on currency. So if it wants to tighten monetary policy, it could *either* reduce the supply of currency *or* increase the demand for currency (by raising interest rates on currency).

    But if the central bank issued a very large supply of currency, at too high a rate of interest, currency would drive out TBills altogether.

  29. Ed writes:

    Tbills have a significant advantage over currency and bank deposits. Unlike currency, there is no risk of theft with Tbills as it can be held electronically. And bank deposits above the $250K FDIC limit aren’t insured and subject to credit risk.

  30. Ed writes:

    Nick,

    Tbills have a significant advantage over currency and bank deposits. Unlike currency, there is no risk of theft with Tbills as it can be held electronically. And bank deposits above the $250K FDIC limit aren’t insured and subject to credit risk.

  31. Philippe writes:

    “So if it wants to tighten monetary policy, it could *either* reduce the supply of currency *or* increase the demand for currency (by raising interest rates on currency).”

    If it raises the interest it pays on reserves, it increases the amount it pays to banks, so the net outflow of money from the Fed increases. Seems odd to describe this as ‘tightening’.

  32. stone writes:

    Do the central banks of the “liquidity trap” countries such as Japan, USA and UK actually still have the capacity to “tighten” if that means increasing interest rates much above inflation? I mean they would have a hard job selling off their holdings of assets for anything like as much as they paid for them so they wouldn’t be able to reduce the monetary base in that way by enough for that to increase interest rates. They wouldn’t be able to pay a lot of interest on reserves for long before that depleted their own source of funds. Am I missing something crucial?

  33. stone writes:

    The Fed now uses reverse repos to provide non-banks with an interest rate similar to the interest on reserves that banks get:
    http://www.bloomberg.com/news/2013-10-14/fed-gets-bigger-in-markets-as-qe-prompts-new-tools.html

    BUT that is way off the Fed having the capacity to induce a substantial interest rate hike as they used to do during the “great moderation”.

  34. Unanimous writes:

    Further expanding on Mark A Sadowski’s points above, central banks can control the total of reserves and currency independently of interest rates. They can set reserve requirements for commercial banks to enable control of interest rates for any particular amount of base money. I only know of one instance of this (Australia during the GFC in which base money was purposely doubled for a period of months), but it could be done at any time.

    To Stone, central banks can’t run out of their own currency.

  35. stone writes:

    unanimous@34, “To Stone, central banks can’t run out of their own currency”

    I thought central banks could create unlimited amounts of their own currency to make purchases with but were not at liberty to create their own currency for the purpose of paying interest on reserves. The interest on reserves has to be paid from funds generated by the central bank’s own portfolio. Obviously these are simply “rules” but in essence I guess such “rules” constitute what money is???

  36. Benjamin Cole writes:

    So..what is the upshot? If the Fed continues with QE, but limits it purchases to US bonds, and then permanently rolls over on its portfolio, has it permanently monetized debt? And thus granted huge tax relief to taxpayers? Lowered out national debt to nationa; GDP ratio? Why don’t we proceed?

    And this will not be expansionary or inflationary?

  37. Unanimous writes:

    Stone@35, yes they buy financial assets, so if they need income to pay interest they can buy financial assets that have an income associated with them (which most financial assets do), this income then pays the interest.

    The “danger” in the QE assets loosing value is that all the QE dollars expended in purchasing the assets can’t be “recovered” by selling the QE assets. Commercial banks would be left with high reserves, and interest rates would be stuck very low. But this isn’t really a danger because reserve requirements can be raised so that commercial banks need the higher reserves left over from the incompletely exited QE, and then the central bank can control interest rates.

    For much of history in many countries there were no central banks and commercial banking functioned not just with zero average reserves as some countries do today, but without even a monetary base capability at all. A “monetary base” isn’t fundamental to banking, the term is misleading. A banking system can function with various amounts of reserves (including negative or zero reserves), and with paying interest on reserves or not, and interest rates can still be under control of the central bank.

    To Benjamin@36: no it is not very expansionary, and not inflationary under present conditions, and it effectively lowers the national debt. The Fed is proceeding. Eventually, this may become common policy throughout the world. It is really nothing different than what the non-government sector has always done with its commercial banks, why shouldn’t government get in on the act too? State governments may also start their own banks too and get in on the act.

  38. stone writes:

    Unanimous@37, thanks for telling me about zero average reserve banking. Do you have a handy link that gives details? In the UK we have no reserve requirement but reserves are used. In the a zero average reserve system you refer to, is there never any settlement? Do banks just end up owing one another various amounts? Don’t banks that end up being owed more and more by other banks get nervous? In the free banking era in the USA am I right in thinking that vault cash was used for settlement and bags of paper cash were carted between banks for settlement. That vault cash would have been the reserves of that era I guess.

    You say that reserve requirements can be used by central banks to give a sufficient value to reserves such as to maintain as high an interest rate as the central bank chooses. I’m just trying to imagine what level of reserve requirement would be needed today in say the USA to induce say a %10 interest rate (not outlandish by 1980s standards).

  39. Jared writes:

    Unanimous – I don’t think raising reserve requirements is a possible solution for the US moving forward. If you raise the reserve requirement to make current excess reserves (almost $2 trillion) become required reserves, then you’ll be strapping banks with billions of zero earning assets. Reserve requirements are effectively a tax on bank assets. You’d be putting banks out of business with any meaningful elevation of the reserve requirement. That’s why increasing IOR is the only way forward once the economy picks up, as Steve argues in this post.

  40. Mark A. Sadowski writes:

    Jared,
    US reserve balances were 13.0% of GDP in 2013Q2. This compares with 13.6% of GNP in 1940Q2 and 1940Q3:

    http://research.stlouisfed.org/fred2/graph/?graph_id=123169&category_id=0

    Back then the Fed didn’t have interest on reserves, only reserve requirements. Interest on reserves is just a different way of doing the same thing.

  41. stone writes:

    Mike@40, I’m wondering how fair a comparison it is between 1940 and today though because then so much more of the monetary base was paper cash in circulation I guess.

  42. stone writes:

    Mark A. Sadowski@40, In 1940, interest rates were also very low weren’t they? Is it possible that reserve requirements back then wouldn’t actually have had the capability to push up interest rates terribly much even had they wanted to (I realize that they wanted low rates at that time)?

  43. […] they have a unique way of central banking. (will it be exported over to the UK~?) From a comment at Interfluidity by Nick Rowe […]

  44. Mark A. Sadowski writes:

    stone,
    Here’s the Friedman and Schwartz M2 money multiplier (in blue, and don’t ask) and the yield on short term (3 to 6 month) U.S. securities from 1925-70 (red):

    http://research.stlouisfed.org/fred2/graph/?graph_id=123223&category_id=0

    Short term rates were under 0.25% until mid-1942 when they were pegged at 0.375%. At that time reserves were about 8% of GNP. After the Fed un-pegged short term rates in mid-1947, they were allowed to rise above 1% in August 1948 for the first time since March 1932. By then reserves had fallen to 6.4% of GDP. During and after WW II they were concerned about holding rates down, and did not appear to have any problem raising rates using reserve requirements.

  45. Mark A. Sadowski writes:

    stone@41,
    “I’m wondering how fair a comparison it is between 1940 and today though because then so much more of the monetary base was paper cash in circulation I guess.”

    In 1940Q3 currency was 37.0% of the monetary base. In 2013Q3 currency was 35.3% of the monetary base. So there’s virtually no difference.

  46. stone writes:

    Mark A. Sadowski@44, thanks for putting me straight on all of this!

    The bit I’m still struggling to grasp is when you say, “During and after WW II they were concerned about holding rates down, and did not appear to have any problem raising rates using reserve requirements”.

    Are you meaning that they could have raised rates a lot had they wanted to but they didn’t want to? They didn’t actually try did they? I’m meaning getting rates up to 1980s levels not 1% rates. By the time they came to raise rates a lot a couple of decades later, inflation and growth had caused the stock of base money to be much smaller relative to the economy hadn’t it (as in the chart from your previous link). Sorry if yet again I’m just displaying my ignorance/confusion.

  47. Jared writes:

    Mark – I’m confused by your conflation of raising reserve requirements and raising rates. The two are quite distinct. The central bank could have a low reserve requirement and a high overnight rate, or a high reserve requirement and a low overnight rate, or vice versa. Raising the reserve requirement only affects the quantity of reserves a bank must hold, which places an extra cost on banks (the opportunity cost of holding other interest-bearing assets), it does not influence the price at which reserves trade.

  48. Mark A. Sadowski writes:

    stone,
    “Are you meaning that they could have raised rates a lot had they wanted to but they didn’t want to? They didn’t actually try did they?”

    During the war the Federal Reserve became a very active purchaser of Treasury debt. The Treasury
    wanted to keep its borrowing costs low and encouraged the Fed to hold down interest rates. This led to the pegging of short term rates in 1942. After the war inflationary pressures emerged. The government ran large budget surpluses but the outstanding debt was still substantial. Thus initially the Treasury resisted Federal Reserve requests to raise interest rates to contain the inflationary pressures. In 1947 the Treasury finally did agree to an upward adjustment of the rates. See pages 33-34 for more information:

    http://research.stlouisfed.org/aggreg/meulendyke.pdf

    “By the time they came to raise rates a lot a couple of decades later, inflation and growth had caused the stock of base money to be much smaller relative to the economy hadn’t it (as in the chart from your previous link).”

    The monetary base remained roughly the same from 1948 until 1959 while nominal GDP grew:

    http://research.stlouisfed.org/fred2/graph/?graph_id=145118&category_id=0

  49. Mark A. Sadowski writes:

    Jared,
    There’s a tradeoff between convenience yield and reserve requirements. Adjusting reserve requirements is one way of making the monetary base arbitrarily large with respect to nominal GDP. Adjusting interest on reserves is just another way of doing the same thing.

    Or, more simply, adjusting reserve requirements and adjusting interest on reserves are just two different ways of adjusting the demand for base money.

  50. […] interfluidity – Why (and when) interest-on-reserves matters… […]

  51. stone writes:

    Mark A. Sadowski@48, thanks for the brilliant link. To be honest though to me it doesn’t seem to give much indication that increasing reserve requirements is such a powerful tool. On page 18 it points out that ” Recent efforts by depository institutions to avoid reserve requirements by sweeping consumer checking account balances into savings accounts have lowered required reserve balances to levels where they are not binding on the behavior of most depositories.”

    From reading that link it looks to me as though the only time that much use was made of greatly increasing reserve requirements was in the late 1930s and then it certainly didn’t crank up interest rates a lot, p31 “. They worried that the excess reserves could set off inflation at some point in the future and consequently sought a way to eliminate them. Open market sales of securities were contemplated, but the excesses were so large that such sales would have reduced Federal Reserve earnings to the point where covering expenses might have been difficult. Discount window borrowing already was negligible, so there was no scope
    for further reductions.
    Instead, the Federal Reserve turned to its new tool, reserve requirement ratios, and raised the ratios dramatically in several steps in late 1936 and early 1937. To the frustration of Fed officials, the banks built up their excess
    reserves again and, in the process, contracted the money stock. At the same time, the Treasury stopped issuing gold certificates to the Federal Reserve against the gold inflows, thus halting reserve injections from that source.
    Economic activity contracted until 1938, when the Fed reduced reserve requirements modestly and the Treasury resumed monetizing gold inflows.”

    Post WWII the use of reserve requirements looks to be very modest p36, “The Board changed reserve requirements occasionally to signal a policy shift. The changes were far smaller in magnitude than those of the 1930s, and the impact on reserves was generally cushioned with open market operations that partially offset the reserve impact.” – and that is talking about the period after the late 1940s period during which the government surpluses and pegged low interest rates had chopped down the level of government debt and inflation and growth had caused the economy to be much bigger in relation to the size of the monetary base which as you say remained the same nominal size through the 1950s.

    To be honest it still looks to me as though the 1980s interest rate hikes might only have become possible after the WWII excess of base money was inflated away, taxed away and grown into????

  52. Mark A. Sadowski writes:

    stone,
    “To be honest though to me it doesn’t seem to give much indication that increasing reserve requirements is such a powerful tool.”

    That may be true, but my point is that having large reserves don’t automatically mean interest on reserves is the only way to deal with the situation.

    “From reading that link it looks to me as though the only time that much use was made of greatly increasing reserve requirements was in the late 1930s and then it certainly didn’t crank up interest rates a lot, p31″

    The 1937 recession was the second most severe US recession of 20th century, with real GNP declining 11% and industrial production falling 30%. This was accomplished without raising short term interest rates much above 0.6%.

    Incidentally, gold sterilization (which stopped the monetary base from growing) was probably a far bigger factor in that recession than the doubling of reserve requirements from 13% to 26%. See Douglas Irwin for more information:

    http://www.dartmouth.edu/~dirwin/1937.pdf

    “To be honest it still looks to me as though the 1980s interest rate hikes might only have become possible after the WWII excess of base money was inflated away, taxed away and grown into????”

    The money multiplier and the velocity of money were extremely low in the US from the early 1930s through the early 1960s, and from 1929 to 1965 PCEPI inflation only averaged about 1.9% annually. Why would one need to raise interest rates to very high levels under such circumstances?

  53. stone writes:

    Mark A. Sadowski@52, I’m not saying that recessions are only induced by high interest rates nor that I’m a fan of using high interest rates to induce recessions. I’m just saying that all this talk of the Fed having the potential capacity to reverse its current minimal interest rate/ flood of liquidity stance may be on shaky foundations. The 1939 to 1948 period did apparently have 8.7% annual inflation according to this link ( http://www.nber.org/chapters/c11389.pdf ). Now obviously war time expediencies meant that such inflation was the least of anyone’s worries. What I’m wondering though is if we were to get 8.7% inflation in the near future, would the Fed (or Bank of England or Bank of Japan) be able to do anything effective to curtail it. I’m still left with the impression that we would need to have a sustained period of such inflation FIRST before the effects of that inflation recreated a 1980 type scenario where the monetary base had been inflated away and grown into to such an extent that the Fed had regained the capacity to put up interest rates.

    My personal priority is to have lots of well paying jobs for everyone. I’m no anti-inflation hawk. I just want to have a clear idea of how our system works (or doesn’t).

  54. stone writes:

    I guess my fear is that they have the policy of doing whatever it takes to keep asset prices aloft and that, as a side effect of that, a scenario is developing where ever more rigorous political efforts are made to keep wages low and unemployment up (as that becomes the only way to ensure “core inflation” is never a danger) whilst at the same time the delusion is put about that monetary policies and tax cuts for the wealthy that actually are only about supporting asset prices are intended to also somehow reduce unemployment and support the real economy????

  55. stone writes:

    I guess my fear is that they have the policy of doing whatever it takes to keep asset prices aloft and that, as a side effect of that, a scenario is developing where ever more rigorous political efforts are made to keep wages low and unemployment up (as that becomes the only way to ensure “core inflation” is never a danger) whilst at the same time the delusion is put about that monetary policies and tax cuts for the wealthy (that actually are only about supporting asset prices) are intended to also somehow reduce unemployment and support the real economy????

  56. Unanimous writes:

    To Stone: the total of reserves plus cash is determined by reserve bank asset purchases. How those reserves are distributed between banks depends on individual bank activities which are influenced by reserve bank policies. Apart from using physical cash, banks can’t control the amount of reserves in the system. Reserves are transaction accounts for commercial banks to use when transferring money with each other, and these transactions net out to zero across the system.

    As well as determining the amount of reserves, the central banks set rules that influence how the reserves are shared between banks. The central bank paying and charging interest on reserve accounts is one way to influence the spread of reserves between banks, and setting an amount for each bank to meet is another. Sometimes banks find ways around the rules. If the rules require the reserves of each bank to be in proportion to particular types of accounts held by the public then banks may minimise the use of these accounts. If the central bank requires banks to hold more reserves than exist, and penalises them for not meeting requirements, then interest rates head up as banks compete for the limited amount of reserves until they are making enough money to compensate them for the penalties they are paying the central bank. If the central bank requires each bank to hold less than their share of reserves and provides no financial incentive to hold their share of reserves, then overnight interbank rates will head close to zero as banks no longer need to compete for the reserves.

    By setting various rules, central banks can set interest rates independantly of the overall level of reserves.

    You asked for an example of a reserve system in which there is an average zero reserve balance. I understand that Canada is close to such a system – the reserve requirement is zero. On days when a bank has a negative reserve balance it pays interest. On days when it has a positive balance it is paid interest (but 0.5% less than the what it pays for a negative balance). Overall the reserves are zero, and for each bank over time they average out close to zero.

    You might ask what is the point of QE if interest rates can be varied indepenantly of reserves. When central banks buy assets from non-banks it also influences monetary aggregates other than base money, and it influences interest rates other than overnight interbank rates. Central banks are trying to influence these other things, which they normally leave alone.

  57. stone writes:

    unanimous @56, what you just wrote pretty much fits in with the overall view of how it all operates that I have. My confusion about your saying “average zero reserve balance” was simply because I would have understood it as zero reserve REQUIREMENT. From what I can see the Bank of Canada system is like the Bank of England system where there is no reserve requirement for each individual bank but overall in the system there is very much a net positive overall reserve balance. The first thing I came to on google was: http://www.pearsoned.ca/highered/divisions/text/mishkin_2/data/appendices/15_ch15_mishkin_append.pdf

    My point is that if that overall stock of reserves in the system gets very large then the central bank may have burnt its bridges in terms of loosing its capacity to raise interest rates by much for any sustained period. The idea that switching to high reserve requirements could rectify that looks doubtful to me. The attempt to do that by the Fed in the late 1930s shows little evidence of having had much effect at all.

  58. Unanimous writes:

    Stone@57. The balance sheet you linked to shows very minimal reserves – M$1,065.6 for the chartered banks. Cash is M$41,146.7. What’s Canada’s GDP – 2 trillion? So reserves are maybe 0.05% of GDP, compared to 25% or something like that for the US. It’s rarely going to be exactly zero at any point in time, but it’s smaller than the error in GDP estimate, so that’s close enough to zero in my mind for the purposes of this discussion. Possibly the central bank views it as unfair or counter productive for the overall reserves to go negative and make commercial banks pay interest to it, so perhaps I was slightly wrong to imply that it would go negative sometimes, although I still wouldn’t rule out the possibility. From the commercial banks’ point of view on any given day – some will be positive others negative.

    As for the other document you quoted on the history of the Fed, I can’t make sense of many statements in that history. A prime example is the following: “Instead, the Federal Reserve turned to its new tool, reserve requirement ratios, and raised the ratios dramatically in several steps in late 1936 and early 1937. To the frustration of Fed officials, the banks built up their excess reserves again.” Why would you be frustrated that banks build up reserves when you have just required that they build up reserves?

    Also, in modern terms banks don’t control reserves as a whole. It’s possible that back in the 30s the word reserves had a slightly different meaning, maybe it was a subset of the total accounts commercial banks held with the central bank. Different terms are used to refer to the transaction accounts commercial banks hold with the central banks – for example in Australia they are called “exchange settlement accounts” – see the following link which also has a graph showing the total exchange settlement funds over time: http://www.rba.gov.au/mkt-operations/dom-mkt-oper.html. Notice that interest rates have varied in quite a different manner to the exchange settlement balance over time, and that the central bank can retain control over interest rates despite the size of reserves varying by a factor of 5. You might also be interested to see in that link that Australia’s government debt is too low for the banking system to function according to international standards without the central bank providing special facilities to enable banks to hold enough zero risk assets. Special facilities like these can also make comparisons a bit tricky when various accounts have different names and conditions.

  59. stone writes:

    Unanimous@58, I’m only wondering about situations where there is a VAST excess of reserves such as currently in the USA, UK and Japan. I still think that creates a different situation. The Fed claims that they have the option of creating certificates of deposit for the commercial banks’ accounts at the Fed so as to encumber reserves if the need arises. To be honest I still think that poses the question of where the Fed would find the necessary interest payments from -just like with paying interest on reserves.

  60. stone writes:

    Perhaps QE is supposed to be irreversible – to lock in very low rates for decades or at least until the reserves get inflated away. Perhaps that is how QE signals to the market? But, if so, why don’t central banks come out and say that they have burnt their bridges. It seems a bit like the secret doomsday device in Dr Stragelove.

  61. Unanimous writes:

    I agree that certificates of deposit don’t solve the problem (the problem being how a central bank can raise rates significantly higher than it receives on its assets when it has a lot of assets). Rules, laws, and regulations that enable the central bank to control rates without paying interest will be the way forward, eventually. Very low rates could be around for a long time, but they will end. I don’t know which rules and laws will be used, but whatever are easiest to change will probably determine which ones get used. My tip is reserve requirements, but it could be something else. The history of banking in allied nations during WWII might be a good place to look for possible precedents. There have been times and places when interest rates (even consumer rates) were decreed without any ensuing disaster. If it is widely agreed that a government needs to control something, then it can.

    Maybe in your terms the bridges are burnt, but how long to build a new one? I don’t think it’s long when the bridge is new rules and there is a pressing need.

  62. stone writes:

    unanimous@61, what happened after WWII was that rates were not put up much until AFTER inflation, taxation and economic growth had caused the stock of reserves to no longer be large relative to the economy. Perhaps that was not a choice, perhaps that is the only way. They hiked up reserve requirements a lot in the late 1930s and it failed to hike up interest rates.

    Perhaps this is what QE is all about, – creating a situation where the Fed can credibly say,

    “don’t hold savings as money in anticipation that rates will go up because they can’t until your money has lost its value -however long that takes.”

    The problem is that lobbying for regulatory and fiscal changes that create a protracted depression is the way that hoards of money can be caused to retain their financial power for long into the future.

    Of course there is a way that government can use fiat to reduce reserves and ensure price stability without paying interest- it is called taxation of the wealthy BUT central banks are not the part of government with that power and politicians (who do have that power) like to think that they have abdicated responsibility over unemployment and price stability to the central bankers.

  63. […] Why (and when) interest-on-reserves matters… -Interfluidity […]