“Some economists”

Robert Hall and Susan Woodward posted an essay on the subject of interest on reserves:

Raising the reserve interest rate is a contractionary measure.  A higher interest rate on reserves makes banks more likely to hold reserves rather than increasing lending. The Fed’s decision to raise the reserve rate from zero to 75 basis points just as the economy entered a sharp contraction in activity is utterly inexplicable. Fortunately, the Fed lowered the reserve rate subsequently, but the continuation of a positive reserve rate in today’s economy is equally inexplicable.  Some economists have proposed that the Fed charge banks for holding reserves, an expansionary policy worth considering. With the Fed funds rate at around 15 basis points, it would take a charge to restore the differential that drives banks to lend rather than hold reserves. Were the Fed to charge for reserves, they would become the hot potatoes that they were in the past, when the reserve rate was zero and the Fed funds rate 4 or 5 percent. Banks would expand lending to try not to hold the hot potatoes and the economy would expand. There is no basis for the claim that the Fed has lost its ability to steer the economy. (However, the Fed would have to go to Congress to get this power, as it did to get the power to pay positive interest on reserves.)  [italics in original]

BTW, I am not suggesting that they are referring to this blog, as I’m sure other economists are also making this argument, but I’d like to think we might be having some impact on the conversation.  I sent an email to Robert Hall a couple weeks ago.

Thanks to Dilip for two very helpful links.  This one, and yesterday’s link to Brad DeLong.  BTW,  Andrew Sullivan just linked to the global warming post.  Make that three, Dilip just sent me the link below.

Update:  Mark Thoma just linked to the Hall and Woodward idea here.  Mark calls their essay “good, but wonkish,” although it’s not clear if that refers to the interest penalty idea.



47 Responses to ““Some economists””

  1. Gravatar of David Stinson David Stinson
    13. April 2009 at 08:09

    Hi Scott. I understand that paying interest on reserves is contractionary in that it reduces the extent to which increases in monetary base result in increases in the money supply. The problem that I have with this issue is that it implies that the build-up of excess bank reserves last fall was a bad thing whereas I had presumed that it was a good thing.
    My assumption was that the build-up of reserves was the most expedient method of increasing the stability of the banking system at a time when raising capital (and in particular, equity capital) was/is problematic for most banks. The events of the last year suggest pretty strongly that while socializing risks provides a facade of stability, removing market discipline ultimately encourages otherwise unsustainable practices, undermines stability and increases systemic risk. In the current environment, it seems unwise to rely too heavily, even in the short run, on non-market-measures, such as deposit insurance or bailouts, to create artificial stability or insulate us from the impacts of instability (such as bank runs or increased loan defaults).
    Given the elimination, through policy-induced moral hazard (such as deposit insurance and Too Big/Interconnected to Fail), of market-based incentives for banks to shore up their stability, one could be forgiven that thinking that the most problematic banks might not increase reserves unless they were given an economic incentive to do so via the payment of interest on reserves.
    The Fed data indicates that most of the accumulation of excess reserves occurred quite quickly, between October and December of last year, and had started to taper off in December. I don’t if I am reading too much into this but perhaps the quick run-up suggests that the accumulation was mostly a one-time event. The build-up in excess reserves did not prevent M2 from growing at around 8 or 9% since September, implying annual growth well into the teens. Also, most of the data one sees indicates that bank lending is actually up.

  2. Gravatar of ssumner ssumner
    13. April 2009 at 08:35

    David, You raise some good points, but there are also some common misconceptions in your comment. First of all, I’m told that T-bills would provide just as much capital as bank reserves, so banks don’t actually need to hold so many excess reserves for the reason you cite. Second, even if I am wrong, and the Fed wants banks to hold several billion in reserves above the normal level, they should still impose a penalty rate on any reserve levels above that which banks need to hold (in the eyes of the Fed.) That would insure that additional OMOs would have traction, would go out into circulation. The key is behavior at the margin.
    The monetary aggregates were completely discredited in the 1980s, as the demand for those assets changes with changing economic conditions. During a financial crisis there is a big increase in the demand for safe, liquid, FDIC-insured assets like bank accounts. So the increase in M2 doesn’t mean anything. Similarly, bank lending, which you say has increased somewhat from a very low level, tells us nothing about expected growth in NGDP–which is still far too low. Until that rate gets up to at least 5%. the Fed is far too contractionary.

  3. Gravatar of David Stinson David Stinson
    13. April 2009 at 09:19

    I can see that T-bills would be (almost) as liquid as monetary base in the event of a run but given the size of the vulnerable institutions and their interconnectedness, I am wondering whether the US government would be keen on one or more banks potentially dumping hundreds of billions of dollars worth of T-bills to raise cash in a hurry.

    I would imagine however that if banks held T-bills rather than excess reserves, the amount of the monetary base (or excess reserves) that we are discussing wouldn’t be available for monetary expansion in any case. So holding T-bills of the same value would be as contractionary as holding excess reserves, wouldn’t it (I’m asking)?

    I recognize that targeting growth in monetary aggregates alone was discredited because it ignores changes in money demand. The only reason that I mentioned M2 growth was to point out that paying interest on reserves had not prevented a significant expansion in the money supply, however defined. I wasn’t suggesting that it was the appropriate rate of growth or that it was a target aggregate of some kind.

  4. Gravatar of David Stinson David Stinson
    13. April 2009 at 09:26

    Sorry. Forgot to mention that I suppose any concern regarding bank stability could be addressed by paying interest only up to specified levels of excess reserves or appling a schedule of interest rates that declined with successively higher levels of excess reserves. The bank stability issue wouldn’t justify paying interest on unlimited levels of excess reserves, presumably.

  5. Gravatar of JimP JimP
    13. April 2009 at 15:01

    Scott –

    Well – far as I am concerned, that Hall Woodward article was a direct endorsement of you and your efforts.

    It is now posted on VOX btw – http://www.voxeu.org/index.php?q=node/3444.

    And I think congratulations are due to you.

    And from the guy who wrote the book on the subject.

    One can only hope that the Fed board will see it.

    In italics!

    From the article:

    “The Fed’s decision to raise the reserve rate from zero to 75 basis points just as the economy entered a sharp contraction in activity is utterly inexplicable. Fortunately, the Fed lowered the reserve rate subsequently, but the continuation of a positive reserve rate in today’s economy is equally inexplicable.”

    Where one can only take the work “inexplicable” as actually meaning, and having been meant by the authors, as “inexcusable”.

    I think it was very low of the authors to not link to your blog.

  6. Gravatar of ssumner ssumner
    13. April 2009 at 15:30

    David, No, if the banks hold less reserves the monetary base does not contract. Banks have no control over the size of the base, the Fed controls the base. Any reserves the banks don’t want goes right into your wallet and mine. Just to put things in perspective, that’s more than $2000 per capita. Are you ready to carry an extra $2000 in your wallet? I thought not. That’s why I’m pretty confident the plan would boost AD.

    With FDIC I don’t expect any big bank runs. And even if there were I really don’t think the Treasury would care if banks suddenly sold a lot of T-bills. Their price is pretty stable, and we are talking about the secondary market.
    I agree with your second point about the foolishness of interest on unlimited excess reserves.

    JimP, I had the same reaction as you–I also thought inexplicable meant inexcusable. Here’s arguably the smartest monetary economist in the world saying there is no justification for the Fed paying interest on reserves, and if anything the rate should be negative. I hope my readers now understand that I am not just some crackpot, and that my occasionally passionate statements of outrage against Fed policy are well thought out and fully justified.

    A mention would have been nice, but the important thing is to get the message out there. And in fairness to Hall he has worked on this issue for a long time, so I suspect he already understood the need for a penalty rate.

  7. Gravatar of JimP JimP
    13. April 2009 at 15:49

    Scott –

    I really really hope Hall is famous enough that this will get through to the Fed. The idea that we would be stranded for years and years in a grinding suffocating deflation – with the national debt rising to the moon to pay for worthless spending – another Japan – just horrifies me.

    These things matter!

    It is not just some abstract and fun debate about the nature of economic theory. We are being destroyed by these polices – and the thought that they could just be changed but just aren’t being changed – it drives me crazy.

  8. Gravatar of JimP JimP
    13. April 2009 at 15:51

    And it drives you crazy too.

  9. Gravatar of JimP JimP
    13. April 2009 at 16:19

    And let me add another thing –

    Far as I am concerned, your blog is the most intelligent analysis of this crisis around.

    Hall’s suggested Fed announcement:

    “The Federal Reserve is fully committed to a policy of stable and low inflation. Though the Fed has not adopted a quantitative target for a specific measure of inflation, its actual performance over the period from 1987 through 2007 is indicative of its goal for the future. The Fed will continue its efforts to expand the economy this year, when inflation appears to be well below its normal range. Its past and planned expansionary policies during the current period of extreme stress will result in a large expansion of reserves. The Fed will use its authority to pay interest on reserves as needed to prevent excessive inflation as the economy recovers.”

    is ok – but kind of sad and passive. It misses the point. The problem now is deflation – not inflation. And it is sad and wordy.

    Your policy prescriptions are sharp, clear, and immediately obvious to anyone with one half of one brain.

    1. Set a price level target – not an inflation rate target – but a price level target – and commit to hitting it.

    2. Stop paying interest on reserves – start charging interest on them. Drive the excess reserves out of the banks as a hot potato to heat up spending and thus employment.

    Two clear and decisive steps.

    Like you, I believe the reaction of both the markets and the economy to these would be very large, and immediate. As you said in the post on DeLong:

    “But the key driver of current changes in AD is expected future changes in NGDP. (I believe this may be part of what Keynes meant by “confidence.”) And this is determined by changes in the expected future supply of money (relative to changes in demand.)”


    Expectations would immediately change and confidence would come back.

    You would go down as the Keynes of this crisis. This is a title Krugman is searching for – of course – and one he in no way deserves. His advice is trash. But you do.

  10. Gravatar of Scott Lawton Scott Lawton
    13. April 2009 at 18:21

    Outstanding news! In addition to the quoted passage, I think this bit is important:

    So the question John Taylor posed-how can the Fed control inflation in coming years when it is committed to have a large volume of reserves outstanding to finance its purchases of illiquid assets?-has a simple and effective answer: The Fed should raise the rate its pays on reserves as needed to control economic activity and inflation.

    Lots of people are nervous about future hyperinflation, so I think it’s important to show HOW a current change to target NGDP (or even just moderate inflation) won’t inevitably lead to “payback” later on.

  11. Gravatar of Dilip Dilip
    13. April 2009 at 18:44

    That thing about inexplicable (or inexcusable), JKH over at Mark Thoma’s blog (with whom Scott has been having a discussion) has a different take on it. Its funny he has quoted this exact same paragraph and reached the opposite conclusion. I am posting a few links to his comments here, but I think Scott should request him to bring his discussion over here. Its easier to follow.

    * http://economistsview.typepad.com/economistsview/2009/04/woodward-and-hall-the-fed-needs-to-make-a-policy-statement.html?cid=6a00d83451b33869e201156f2259f3970c#comment-6a00d83451b33869e201156f2259f3970c

    * http://economistsview.typepad.com/economistsview/2009/04/woodward-and-hall-the-fed-needs-to-make-a-policy-statement.html?cid=6a00d83451b33869e201156f225fe6970c#comment-6a00d83451b33869e201156f225fe6970c

    * http://economistsview.typepad.com/economistsview/2009/04/woodward-and-hall-the-fed-needs-to-make-a-policy-statement.html?cid=6a00d83451b33869e2011570194217970b#comment-6a00d83451b33869e2011570194217970b

  12. Gravatar of JKH JKH
    14. April 2009 at 01:18


    I don’t know if you’ve seen this article “Divorcing Money from Monetary Policy”


    It’s a description of the functional relationship between the fed funds rate and the interest rate paid on reserves.

    Such a functional relationship only exists of course when the Fed actually implements a structural policy of paying interest on reserves. The Fed only started such a policy last September. So the relationship didn’t apply prior to that.

    The Fed moved to a policy of paying interest on reserves (or more precisely, on excess clearing balance levels held at the Fed) because the level of excess balances itself became structurally and historically excessive. The Fed created the outsized excess position deliberately because it wanted to use excess clearing balances for the first time as a material source of funding for its extraordinary balance sheet expansion. Such balance sheet expansion and such use of clearing balances have not occurred historically. But because such outsized levels of excess clearing balances would otherwise threaten the Fed’s control over the effective fed funds rate (i.e. the floor of the trading range), it became necessary to pay interest on balances in order to maintain control. This was definitely the case when the Fed changed to this policy last fall, because the target fed funds rate was still well above zero.

    That said, having the fed funds target so near the zero bound now is a special case, where arguably the interest rate paid might be zero instead of 25 basis points (and indeed it might be negative as you propose). Today’s zero bound case was also not the case when the Fed implemented its policy last September, because the target funds rate was still well above zero. But the general case remains a structural issue because the level of excess balances is so abnormal relative to history. It is the general case that is important, because the Fed wants the flexibility to change the fed funds target rate in the future, without necessarily having to fully wind down its asset programs prior to making such a change and without necessarily reducing excess clearing balance levels back to their normal, historic, pre-crisis order of magnitude, at least not all at once.

    The historic level of excess clearing balances held at the Fed has generally been low enough that it has been binding at the margin without the Fed paying interest on excess balances – i.e. binding in terms of forcing the banks to compete for relatively scarce clearing balances in order to meet their individual requirements, so that the Fed can control the downside of the effective fed funds range without paying interest on reserves. This is no longer the case due to the outsized levels of excess clearing balances.

    The Fed is now paying the rate of interest on reserves it deems necessary in order to maintain a 0 to 25 basis point range on fed funds. (I believe it is the case that not every institution maintaining balances at the Fed is actually paid interest, according to the rules for various categories of institutions. That may account in part for why the effective fed funds rate would trade below the interest rate on reserves at this time.) In any event, it appears the Fed is being cautious at the zero bound point and doesn’t want to experiment with the effect of either zero or negative rates on either fed funds or reserves. Still, I believe your criticism of the Fed for not charging for reserves should really be directed to why they don’t lower the fed funds target range into negative territory, which would then force a negative interest rate on reserves in order to be structurally consistent at an operational level. This logic also applies to you criticism of why they started paying interest last fall. Essentially, they were forced to because they had not yet decided to lower the target fed funds range to zero. Criticism to the degree it’s directed to reserve interest should be directed first to the question of why they hadn’t already reached the zero bound on the target funds range.

  13. Gravatar of JKH JKH
    14. April 2009 at 02:48

    I’ve noticed a tendency on this and other economic blogs to some confusion on the nature of bank capital.

    Capital is not an asset. Simplified, it is the difference between assets and liabilities.

    Capital can only be created by issuing it new, or by generating it through retained earnings. Either way, capital is a source of funding that can be invested in assets, but it is not an asset.

    Specifically, treasury bills aren’t capital. Somebody has written somewhere that “adding treasury bills” or “adding reserves” increases capital, or that the two are alternative sources of capital. This is nonsense.

    The purpose of capital is to support risk taking and absorb losses when they result from risk taking. Therefore, banks require capital to assume risk. That’s the whole purpose of regulatory capital adequacy standards.

    Therefore, banks require capital to assume credit risk in new lending.

    Excess reserves can only motivate banks to lend to the degree they already have the capital position to support the additional credit risk resulting from such lending. Excess reserves typically motivate banks to undertake short term money market lending because excess reserves are typically temporary, and money market lending is typically low or near zero risk in the case of treasury bills.

    The entire credit crisis has its operational beginnings in the failure of the interbank lending market. This was due to the perception of much higher risk than usual for bank credits in general. Potential lenders were constrained not only by their perceptions of bank counterparty credit risk, but because of constraints on their own capital due to credit problems on their own balance sheets. Excess reserves were hoarded in this environment, more so than normal. But the hoarding was primarily a function of credit risk and capital considerations. Paying interest on reserves at the lower bound of the fed funds target range isn’t going to change this primary dynamic of credit risk and capital constraints.

  14. Gravatar of JKH JKH
    14. April 2009 at 03:10

    Re 2nd last paragraph above:

    “Typically temporary” in this context means prior to the credit crisis, the outsized increase in excess reserve levels, and the payment of interest on reserves. But whether you consider current excess reserve levels longer than temporary, or permanent, doesn’t change the underlying constraint, which is that capital is required in order to assume credit risk of any magnitude, notwithstanding excess reserves. Low risk money market instruments can be taken on, perhaps for a longer period now, where they don’t impinge significantly on capital constraints. But the Fed will still set the lower bound for the fed funds rate via the reserve interest rate, which will affect lower bounds for rates on short term money market instruments as well, and therefore the extent to which banks will accumulate low risk money market assets in an attempt to deploy their individual excess reserve positions.

  15. Gravatar of JKH JKH
    14. April 2009 at 03:29

    I’ve noted the fallacious nature of the textbook “reserve multiplier” effect enough times now. Another way of stating it is that the multiplier theory has been completely superseded by two events/facts:

    a) The nature of capital as a prerequisite for risk taking. It is capital that determines risky lending, not excess reserves. See above.

    b) The fact that central banks up until last September responded to banks’ required reserves by simply supplying them. Thus, the causality flowed from lending to deposit creation to reserve requirements to reserve supply. Central banks then calibrated this macro money process by guiding it indirectly via the policy rate. But the supply of excess reserves (i.e. excess clearing balances at the Fed) was a fairly steady time series over the long run. Starting last September, the Fed began to supply required reserves, and much, much more. The reason is described above – a substantial change in the use of reserves for the Fed’s own financial intermediation, due to the Fed’s extraordinary asset programs and balance sheet expansion. This change then required a further change in terms of the payment of interest on reserves so that the Fed could continue to retain control over the policy rate as its primary lever for monetary policy.

  16. Gravatar of Alex Alex
    14. April 2009 at 04:34


    “…Just to put things in perspective, that’s more than $2000 per capita. Are you ready to carry an extra $2000 in your wallet? I thought not. That’s why I’m pretty confident the plan would boost AD…”

    Have the Treasury mail a check for $2,000 to every man, woman, and child. Finance the tax rebate with new debt. Let the Fed buy the new debt. Don’t worry about raising future taxes. If that doesn’t boost your AD then nothing will. Although lets hope that the rest of the Americans are not like me. I do have $2,000 in cash, crispy uncirculated $100 dollar bills with sequential serial numbers. I remember reading some articles a while back in grad school about how in the near future the only people demanding cash would be drug dealers … looks like they forgot about Argentinians.

  17. Gravatar of Joe Calhoun Joe Calhoun
    14. April 2009 at 05:54


    By the way, I watch the TIPS market as my market indicator of inflation expectations. The spread right now is roughly 1.3% on the ten year. What should be the Fed be targeting? 2%?

  18. Gravatar of Joe Calhoun Joe Calhoun
    14. April 2009 at 05:57

    Sorry Scott, that last comment was for the previous post.

  19. Gravatar of David Stinson David Stinson
    14. April 2009 at 06:02

    Hi JKH.

    Interesting discussion, thanks. FWIW, I think the Bank of Canada makes a practice of paying interest on positive balances.


    Regarding your point about bank capital vs. reserves, the previous discussion above was not intended to convey the notion that reserves were a source of bank capital or owners equity. Rather, the notion was that reserves and capital can both contribute to the perception of bank stability in the minds of depositors.

  20. Gravatar of JKH JKH
    14. April 2009 at 06:40

    Thanks, David – good paper from the Bank of Canada:

    “More specifically, the Bank is the ultimate supplier of settlement balances to institutions in a deficit settlement position, and the Bank accepts deposits of excess settlement balances from institutions in a surplus position. As a result, the Bank can set the prices of such loans and deposits, and thus effectively control the overnight interest rate.”

    That pretty well says it all. The priority is controlling the overnight rate (Canadian counterpart to the target fed funds rate). That’s why they pay interest on surplus reserve positions. They’re not at zero yet, and I don’t expect them to abandon their basic architectural concept if/when they do get to zero.

  21. Gravatar of JKH JKH
    14. April 2009 at 07:16

    Congratulations to Barack Obama for actually referring to the idea of a bank multiplier correctly in his speech today – i.e. as a capital multiplier rather than a reserve multiplier (with reference to TARP).

  22. Gravatar of travis travis
    14. April 2009 at 07:56


    You’ve used the metaphor of a ship (NGDP) in a channel buffeted by winds (shocks) and the Fed’s job is to steer against the winds. In the recent financial crisis, I think a good argument could be made that the winds were the weakness of the financial system due to overleveraging on houses. I wonder what you think were the winds buffeting NGDP in 1929?

  23. Gravatar of Jon Jon
    14. April 2009 at 08:07

    “Specifically, treasury bills aren’t capital. Somebody has written somewhere that “adding treasury bills” or “adding reserves” increases capital, or that the two are alternative sources of capital. This is nonsense.”

    JKH is exactly right. I tried to make this point before, but apparently it did not sink in.

    “The entire credit crisis has its operational beginnings in the failure of the interbank lending market.”

    But this bit is nonsense (for US institutions). There has been no appreciable change in interbank lending among Domestic institutions, nor has the Fed had any trouble suppressing the FF rate.*,**

    Banks doing dollar business but without direct access to the Fed Funds market (+ the Fed itself) experienced elevated interest-rates (as seen for instance in the Libor(s)). This is a reflection that these institutions rely on the short-term paper markets rather than the FFs market or Fed auctions. Across the board, bank CP rates have carried a risk premium reflecting the “publics” distrust of banks, not banks distrust of each other., **

    * The dip in interbank lending does not coincide with the September/October panic. It coincides with the flood of excess reserves: http://lostdollars.org/static/interbank.png

    ** The actual Fed Funds rate was always below the target. Non-bank CP markets remained full functional through the crisis and carried only slightly elevated risk premia. Bank paper carried a substantial risk factor. The curve given is in aggregate. Domestic institutions had substantially lower premiums. Foreign banks had premiums closer to the Libor rate.

  24. Gravatar of Jon Jon
    14. April 2009 at 08:15

    “Potential lenders were constrained not only by their perceptions of bank counterparty credit risk, but because of constraints on their own capital due to credit problems on their own balance sheets.”

    The loan officer survey data suggests that this is not true. In particular, respondents indicated that neither reserve adequacy nor capital adequacy was a control factor in the declining lending volumes. They attributed as the primary cause: a reduced demand for credit.

    US Banks typically have close to 1.5-2x as much capital as required by law. I find it doubtful then that capital adequacy directly regulates money creation. Although the principle could be valid at certain times and for certain institutions. The one exception that I can think of is that banks perceive loan-loss reserves as too small for the actual risks. Given that the balance between loan-loss provisioning and retained earners is partially an accounting/income-tax issue, I could see how the effect capital cushion banks require could be higher than the legal limit.

  25. Gravatar of travis travis
    14. April 2009 at 08:33


    I think that you are not accounting for the fact that paying interest on excess reserves is contractionary. Banks have a choice between several asset classes: cash, reserves with the Fed, short term debt (t-bills, commercial paper, etc), long term debt (loans, govt and corporate bonds). If the Fed didn’t pay interest on reserves, it is possible that banks would still keep the same amount of reserves with the Fed, but it is *much more* likely that *some* of the reserves would be switched into t-bills and commercial paper. That would lower the interest rate on short term debt that would over time translate into lower interest rates on long term debt.

    The Fed is paying interest on reserves for a simple reason: it wants to print money to buy assets without needing the Treasury to sterilize the operations by selling debt and then depositing the cash with the Fed. In essence, paying interest on reserves *sterilize* Fed interventions without needing to call the Treasury.

    If the Fed stopped paying interest on reserves *and* didn’t unwind its asset purchases, I’d bet a lot of money that inflation would explode through the roof. The reason is that reserves are so great right now.

    Also, by paying interest on reserves, the Fed can pursue a contractionary monetary policy if/when the economy starts to rebound without having to sell assets.

  26. Gravatar of Mark Mark
    14. April 2009 at 09:26

    Scott- I’m a former grad student at Bentley, from the MSF program. Thanks for putting our school “on the map”.

    Anyway, I’d like to point something out and see if you wouldn’t mind commenting. Throughout much of 2008 and into early 2009 the Credit Default Swaps that trade on the US Soverign debt rose drastically. This was consistent with the deflation argument, though I’m not sure many noticed. From the way I understand it, when assets and income are falling and real debt is rising, there is a greater risk of a country defaulting on it’s debt. Indeed if the “deflationary spiral” were to continue (or actually just show up, since all we’ve had is disinflation) it is correct to assume there is a greater chance of a country would be forced to default.

    Likewise, when the CDS declines, one would expect that to signify a pickup of inflationary pressures (or expectations) since a country experiencing inflation would have no default concerns. Instead, the nominal level of the debt remains the same and the nominal level of assets and income will rise.

    I’m not sure if I’ve provided a clear example, but I’m sure you get the picture. Which brings me to the question. While deflation was a concern in 2008, and CDS rates on US sovereign debt rose from like 10bps to 100bps. What can we make of the situation now, given that the same rates dropped from 100bps to just over 40bps, and at a much faster pace (from mid February to today)?

  27. Gravatar of JKH JKH
    14. April 2009 at 10:37

    Jon – I was thinking of Libor, which is a massive dollar market. You’re right; it’s wrong to generalize that to domestic fed funds. The point on capital is generic and factual; banks require capital to take risk of any type, including credit risk. And capital constraints were an obvious reality in the crisis, with various banks lining up serially for new equity issues.

    Travis – I agree paying interest on reserves is contractionary relative not paying interest. And I agree it’s a form of sterilization. The combined effect of outsized excess reserves plus interest paid on those reserves is close to neutral when compared to the alternative of normal (low) levels of excess reserves without interest. I agree there may be more asset purchases without paying interest, but banks are still constrained by credit risk and their own capital positions in the quality of the assets they may buy as a result. And the Fed is constrained operationally in wanting a lower bound for the behaviour of the funds rate relative to target. And I agree with your final point that paying interest on reserves allows the Fed more flexibility in timing the unwinding of its asset programs while tightening policy at the same time.

  28. Gravatar of Jon Jon
    14. April 2009 at 11:32

    “And capital constraints were an obvious reality in the crisis, with various banks lining up serially for new equity issues.”

    As I understand, the first round of banks were threatened that if they didn’t take the deal now they’d be cut loose if they ever got in trouble. One of the ironies of the program is that you cannot qualify for funds if your capital ratios are already too close to the limits.

    Let me quote the Loan Officer survey summary:
    “only about 25 percent of the domestic respondents that had tightened standards or terms noted that a deterioration in their bank’s current or expected capital position had contributed to the change… On net, about 60 percent of domestic respondents reported a reduction in demand for such loans from firms of all sizes”

  29. Gravatar of JKH JKH
    14. April 2009 at 12:12

    I’m thinking of pre-TARP, non-TARP capital issues; i.e. market issues.

    But your point on the loan survey is very pertinent. No doubt demand is a big factor. And there seems to be quite a debate as to the ongoing accurate measurement of credit availability, credit demand, and credit activity. All the players have their own story on it. I’m not denying demand is a big factor; that’s to be expected in recession/ depression like conditions; just making an “other things equal” point about capital adequacy.

  30. Gravatar of ssumner ssumner
    14. April 2009 at 16:46

    Wow, a lot of comments tonight. I will break this up into bite size pieces:

    JimP, Thanks. Let me point out that I don’t have any idea who will emerge as “the Keynes of this crisis.” There is only one thing I am sure of; it won’t be someone parroting ideas from obsolete books written 73 years ago. It won’t be someone recycling anecdotes about burying money in the desert and digging it up again. The new Keynes will have new ideas for the new problems we face.

    Scott, Thanks, I agree.

    Dilip, Thanks.

    JHK#1, Yes, I agree about the relationship between interest on reserves and the fed funds rate during normal times. But it also breaks down when the interest on reserves is negative. T-bond yields cannot go sharply negative, because of the cash alternative (which banks don’t have.) I’ve read the NYFed piece, and I agree on their explanation of the Fed’s reasoning. The desire to prevent lower interest rates is what Hall and Woodward called a “confession” of contractionary intent. You are partly correct in that my real criticism back then was partly based on me objecting to their interest rate policy. But there is more, and interested readers should wake up here because I have a different take that I haven’t seen anyone make before (the following should have been a new post):

    The harm from interest on reserves is much greater than one would imagine from merely looking at the impact on interest rates. That’s the Keynesian view, which doesn’t come close to capturing the full picture. Remember that what really matters is changes in the future expected path of monetary policy. In early October I think markets may not have initially understood what to make of the Fed policy, but they quickly learned that in this new world, no amount of QE would make any difference, as interest on reserves would assure that it was all hoarded once T-bill yields fell below the reserve rate. This was a very contractionary shock for the economy, it meant that even future QE was unlikely to work. As this gradually dawned on the stock and commodity markets, prices collapsed. Real estate markets saw what was happening, and the real estate downturn intensified.

  31. Gravatar of ssumner ssumner
    14. April 2009 at 17:12

    JKH#2 I know they aren’t literally capital. If I made that comment it was in response to someone saying they wanted to hold reserves to build up their capital. But T-bills also count 100% in any calculation of net capital position, or at least so I’ve been told. Riskier assets are valued at less than 100%.

    JKH#3 Banks are not hoarding reserves because of risk, but because of opportunity cost issues. If there was no interest and yet T-bills were earning 5%, Excess reserve demand would fall by 99%, banking crisis or not.

    JKH#4, I agree the textbook multiplier presentation is very misleading, very mechanical, it’s not the best way to think about money and banking issues. BTW, the Fed doesn’t “supply required reserves.” They don’t even supply reserves. They supply base money, the public determines how that money is allocated.

    I fully agree that if the Fed wanted lots of ERs for some reason, they’d have to pay interest to get them if they didn’t want to lose control of their target rate. Hall/Woodward and I cannot understand why they didn’t want to lose control of their target rate.

    Alex, I appreciate the goal you have, but we can do this without a lot of added debt, which is the last thing we need.

    Joe, 2% is about right, but look at the 5 year spread, which is only around 0.8% or 0.9%. That one needs to be much higher.

    David, I agree.

    JKH#5,6: I agree. BTW, for Canada the exchange rate is the key. Without adjusting that they are tied to our deflationary policies. And even with it, they are greatly affected by our economy.

    Travis, Good question. 1929 was much different. There were no winds at all. The sea was as calm as it has ever been in American history. And the Fed decide to turn the wheel sharply toward deflation to stop the stock boom. They greatly increased the gold ratio. Hoarding of gold was deflationary under the gold standard. And now the Fed is again encouraging the hoarding of the medium of account–making money more valuable.

  32. Gravatar of ssumner ssumner
    14. April 2009 at 17:33

    Jon, Thanks for the valuable info on banking, I have learned a lot.

    JKH: Thanks also to you. Although I often disagree with you on monetary policy, I learn a lot from you and Jon about banking.

    Travis, I mostly agree, although I would add that things have gotten so depressed that even a zero rate might not be enough, which is why I prefer the interest penalty as a more certain way of turning expectations around.

    Mark, I wish I knew the answer. Maybe someone more knowledgeable can help. I was always taught that default on Treasury debt was absolutely impossible, as the government would run the printing presses to pay it off before defaulting. I still think that. But I suppose anything is possible. Maybe if the burden becomes too great, and yet the public is too inflation-phobic to allow that option, there might be a default. I think 40 basis points is still high, but as to why it fell I suppose the risk of an outright depression has receded. It now looks like a very bad recession, perhaps very long lasting. But things look a bit better than even a month ago.
    BTW, I hope you and any other alums reading this blog will talk it up at school events. Bentley officials know about it, but I don’t think (outside the econ department) they are aware that we are now part of the ongoing conversation about policy at a pretty high level–interacting more and more frequently with other top bloggers.
    Thanks to commenters both in and out of the Bentley community for making the blog better. I know far more about these issues than when I started.

  33. Gravatar of Jon Jon
    14. April 2009 at 18:55

    “I’m thinking of pre-TARP, non-TARP capital issues”

    Fair enough. The Fed’s loan officer survey is monthly, so its difficult to single out the post-crisis, pre-TARP interval. Perhaps there is other data elsewhere.

    If I understand correctly though, you were suggesting that capital regulates credit growth normally. Or do you only mean that capital regulates credit growth during a crisis such as the present one?

  34. Gravatar of Jon Jon
    14. April 2009 at 19:31

    “The fact that central banks up until last September responded to banks’ required reserves by simply supplying them. Thus, the causality flowed from lending to deposit creation to reserve requirements to reserve supply.”

    But that statement is not dispositive in support of your prior remark, “Another way of stating it is that the multiplier theory has been completely superseded by two events/facts:”

    The Fed and other CBs no longer attempt to link money growth + GDP directly. Rather the goal is to be flexible about the exact notional amounts required, but this is not the same as saying that the base is irrelevant.

    Nor does focus on interest-rates as the embodiment of policy imply that base isn’t the lynch-pin of the dynamics.

    The amount of currency leaking into reserves depends on the relative mismatch between creation and demand. The Fed could look at that notional mismatch, but history instead has guided them to look at the interest-rates–really this is nothing but a normalization. So the Fed picks a target-rate. They create base. They monitor the FF market to determine whether to increase or decrease the rate of base creation [to achieve their target]].

    Nothing about this answers whether the base or the interest-rate is the operative component of policy. It could be one or other, policy would be effective either way, but the relative effectiveness could depend on the structure of the OMO. In particular, if short-rates influence GDP/inflation significantly on their own, then OMO of short-instruments such as tbills is ideal. But if its the ultimate quantities that influence GDP/inflation then OMO of notes or bonds is superior (although both work in either case).

  35. Gravatar of JKH JKH
    14. April 2009 at 20:05

    It’s not difficult to identify the pre-TARP period in terms of the capital injection version of the TARP, which only started last fall. There were plenty of bank equity issues before then. I’m suggesting capital adequacy is a generic issue that guides credit and other risk allocation in all environments.

  36. Gravatar of JKH JKH
    14. April 2009 at 20:12

    “Another way of stating it is that the multiplier theory has been completely superseded by two events/facts:”

    I haven’t contradicted this at all. The correct causality I noted is the reverse of the incorrect causality described in textbooks. Textbooks argue that banks lend and create deposits as a function of available reserves. This is wrong. The correct causality is that the Fed supplies reserves in response to requirements generated by loan and associated deposit growth, and then calibrates its desired path for the economy including these financial conditions by setting the policy interest rate.

  37. Gravatar of Jon Jon
    14. April 2009 at 21:12

    But you haven’t shown that the policy-rate itself guides the economy. By subordinating the creation of base in reflexive terms you imply that it is not the base creation that guides the economy; you seem to treat it as incidental. Perhaps this isn’t what you mean to say, but your choice of wording pushes my buttons on this issue.

    That the Fed wants to create the Goldilocks ‘just the right amount’ of money in keeping with growth+preferences is hardly in dispute. They do treat the details of the magnitude as free variable. This is the lesson of our experiences with direct monetarist M2/3 targeting based on dGDP=df(M).

    But is the short-rate or the quantity the hand that guides the economy? Keeping in mind that we all agree that which ever does the work, the Fed uses the FF rate as discipline.

  38. Gravatar of Jon Jon
    14. April 2009 at 21:32

    “The correct causality is that the Fed supplies reserves in response to requirements generated by loan and associated deposit growth”

    And if the Fed fails to do this? The next round of expansion is predicated on the Fed’s response to the first even in your model. Therefore, I could just as well say that reserve creation precedes lending.

    The multiplier equation is an accounting identity:
    M reserves becomes critical and expansion necessary.

    I don’t think the textbooks are wrong. They are just a toy–perhaps we agree.

  39. Gravatar of Bill Woolsey Bill Woolsey
    15. April 2009 at 02:13

    If we imagine that the money multliper rises back to normal and the supply of bank credit and deposits rise to reflect the current value of the monetary base, then contracting base money enough to reverse this would be painful. First the money supply doubles. And then, after it doubles, it is cut back in half.

    If we further imagine that the price level were to rise enough to reflect this increase in the money supply (and, while we are at it, velocity rising back to normal) then contracting base money to reverse all of this would be even more painful.

    However, I just can’t understand Rogoff’s point. The Fed gets weekly information on bank reserves, credit, and deposits. It is hard for me to imagine that one week the Fed will discover that these quantities have all increased a vast amount. So, for the most part, what will be happening is that the Fed will reduce base money while deposits and bank credit are rising. It will just dampen the possible increase. Sure, they could overcompensate (undershoot.) But the notion that having the Fed increase base money to avoid a credit contraction implies a credit contraction when they reverse course.. is wrong.

    Because data on velocity come very late (basically with quarterly GDP) the ability to contract the quantity of money as velocity rises will be more problematic.

    One issue that is sometimes made more or less explicitly is that this time, the Fed is lending to weak banks or into weak sectors of the credit market. They are directing credit. Pulling that money out may be painful, especially for the banks involved. Or, for example, people who can only borrow if the lenders can sell the loans to bundlers who, with some mix of Federal Reserve and treasury money, can sell securitized loans.

    I was speaking to the chairman of BB&T. He was claiming that good credit risks can get car loans, they just don’t want to buy cars. Poor credit risks need to have the Fed and the Treasury support securitization.

  40. Gravatar of JKH JKH
    15. April 2009 at 03:29

    ” By subordinating the creation of base in reflexive terms you imply that it is not the base creation that guides the economy; you seem to treat it as incidental.”

    Absolutely, the base is incidental. That’s an abrupt and slightly exaggerated way of putting it, but it’s close enough to the truth that it’s worth saying and repeating over and over.

    A very general comment I would make about all of the discussion on this blog and probably elsewhere is that people are failing to partition the monetary universe into two parts.

    The first part is prior to the Fed starting to run up its extraordinary levels of excess reserves – i.e. prior to last September or thereabouts.

    The second part is since then.

    As I’ve noted previously, part II uses the monetary base in an entirely different way than part I. It’s using it to fund a substantially increased level of financial intermediation via the Fed. But using it as Fed funding is an entirely different idea than expecting some explosion of monetary velocity as a resulting of the banking system’s asset use of the same money.

    You can’t conflate part I and part II in the analysis, but everybody’s doing it.

    Part I and Part II can be differentiated according to the level of excess reserves in the form of excess deposit balances with the Fed.

    Call Part I excess reserve deposit balances E.

    Call Part II excess reserve deposit balances E + F.

    In part I, the base is unambiguously incidental to monetary policy. That’s because the Fed manipulates E in order to control the funds rate. E is a tiny part of the monetary base. And E is operational; not strategic. It’s a tool for controlling the Fed funds rate. The rest of the base is incidental. If the Fed thinks the public’s demand for currency is growing too fast, or a thousand other things are happening too fast, it will raise the fed funds target. It will announce, and apply a tiny bit of pressure on the supply of E in order to effect the change. E is the fine tuning mechanism. But E is tiny. Less than $10 billion in a monetary system that is several thousand times that size.

    In part II, the base is incidental in a different way. F as part of the base is now very substantial and growing aggressively. And F is required to fund the Fed’s asset programs. But the Fed “sterilizes” F by paying interest on it. In that restricted sense, I can still say that F is incidental to the core monetary policy of controlling the funds rate. In fact, the reason for paying interest is to make it incidental.

    Obviously, we live in extraordinary times, and F is not incidental in the sense that it is a critical part of the Fed’s extraordinary intermediation activity. That’s understood. But that’s why it’s important to be precise about differentiating between Part I and Part II and between E and E + F, and the relationship in terms of the core objective which remains controlling the funds rate.

    When all of this extraordinary Part II stuff is unwound (hopefully), we then go to Part III = Part I more or less, and we return to E where the base is unambiguously incidental to policy.

  41. Gravatar of JKH JKH
    15. April 2009 at 03:44

    “And if the Fed fails to do this? The next round of expansion is predicated on the Fed’s response to the first even in your model. Therefore, I could just as well say that reserve creation precedes lending.”

    The Fed will only “fail” to do this via a relatively minute tapping on the brakes of E as defined above for Part I. This would only be a marginal incidental mechanism in implementing a decision to increase the funds rate. And the Fed may have a thousand reasons for wanting to increase the funds rate, not just that it dislikes the pace of credit or deposit expansion.

    The “next round of expansion” does not occur as a multiplier effect. It occurs as a 1:1 correspondence between aggregate loan expansion and aggregate deposit expansion (more or less; other balance sheet items are involved obviously).

    The commercial banking decision process in this context is not lending from reserves. Banks employ asset liability management to square up their individual reserve positions on a daily basis. The bank lending decision would occur in the context of knowing that the Fed may be in tightening (higher rates) mode for a thousand reasons; not that the bank may be unable to square up its reserve position as a result of marginal lending.

    These are not easy ideas to communicate. But I think the problem with this multiplier thing is that people who believe in it are treating it as causally strategic. It’s not. It’s an identity, and a useless one at that, because it gets the underlying causality (credit expansion creates money expansion) completely wrong, and therefore it gets the dynamics of 1st and 2nd derivative changes in bank balance sheets completely wrong. And it completely ignores the truly instrumental role of bank capital allocation in guiding the path of bank balance sheets. The deposit multiplier is a Stone Age relic of (mis)understanding of how a fiat money system operates.

  42. Gravatar of JKH JKH
    15. April 2009 at 03:54

    “The Fed will only “fail” to do this via a relatively minute tapping on the brakes of E as defined above for Part I.”

    The Part II operating mode is different, of course. The Fed merely announces both the target funds rate increase and the new higher rate to be paid on excess balances. This sets a new range and a new floor for the funds rate.

    Drive on (for several years maybe) until all of F is drained from the system, and the Fed reverts to Part I operations in terms of using a very small amount of E to effect a tapping on the brakes (rate increases) or pushing on the accelerator (rate declines).

  43. Gravatar of Jon Jon
    15. April 2009 at 08:32

    JKH: Lets restrict ourselves to discussing normal times. Yes, the Fed is doing something differently now. This is really irrelevant to the point you are trying to make.

    In the past few years the rate of monetization has been about 4% of base. Historically (1984-2004) it swung between 5 and 10%. This is not an incremental nudge as you suggest.

    You seem to be speaking through my point, which is that the Fed can build an operational model based on monitoring and regulating short-rates, but that does not mean the effect of monetary policy is (primarily) transmitted through short-rates.

  44. Gravatar of JKH JKH
    15. April 2009 at 11:15

    There are two components to the monetary base; bank clearing balances at the Fed, and currency. Each has a very different role in monetary policy. The rates of monetization you cite make sense. Such monetization occurs almost entirely though the issuance of currency; not through clearing balances (in normal times). This is an important distinction.

    The cumulative growth in bank clearing balances held at the Fed has been negligible in normal times. But this is the channel through which the management of the fed funds rate occurs; not through currency. Clearing balances are where the incremental nudge or withdrawal happens when the Fed wants to lower or increase rates. The reason the balances are so small and the effective fed funds rate so sensitive is that the Fed induces its desired interest rate response by forcing the banks to compete for scarce clearing balances in settling their daily positions. Increase the scarcity and the rate moves marginally higher. Decrease the scarcity and it moves lower. Not only is the clearing balance (normally) negligible in macro terms (about $ 10 billion), but the effective fed funds rate is hypersensitive to changes in it and changes in the distribution among banks and therefore the distribution of supply and demand among individual banks.

    In terms of currency monetization, which is the larger component, it obviously makes a difference whether trend growth is 0 per cent, 5 per cent, or 20 per cent. I suppose you could analogize currency monetization to a car and the Fed’s management of clearing balances as the steering wheel, accelerator, and brakes. The currency monetization trend will reflect the trend of real growth and inflation in the economy. This is the result of monetary policy, like the route travelled by the car is the result of steering it and controlling its speed. The steering wheel and speed for monetary policy comes from the Fed’s actions with respect to the relatively small component of bank clearing balances, according to the intended effect on interest rates. If the Fed wants to slow the car down, it will put on the brakes by raising the fed funds rate. This requires only an announcement effect, plus usually the smallest degree of marginal restriction on clearing balances. This will affect economic activity and inflation, which will affect the public’s demand for currency and therefore the Fed’s requirement for monetization using currency.

    Policy is transmitted through short rates. This transmission then affects everything else – the yield curve (based on expectations of the short rate), credit demand (and therefore broad money supply), economic growth, inflation, and yes, even demand for currency.

    The clearing balance component and the currency component of the monetary base are entirely different. Monetary economists should be acknowledging this far more than they do. (They don’t as far as I can tell.)

  45. Gravatar of ssumner ssumner
    15. April 2009 at 17:48

    I do agree with JKH on one point, in the very short run the current operating procedure of most central banks is to target interest rates and let the monetary base adjust passively. Even though this makes it seem like interest rates are the essence of monetary policy, I think this view is wrong. For the following reasons I think the monetary base is the fundamental force affecting the economy, and the fed funds rate is of trivial importance:

    1. Contrast a policy of pegging the interest rate at a fixed level, and pegging the money supply at a fixed level. Neither are ideal, but at least with a fixed money supply the price level is not indeterminate. With a fixed interest rate the price level becomes indeterminate.

    How can this be? I think the answer is that Taylor rule-type policies disguise what is really going on. When inflation rises above target, the Fed might raise the fed funds rate. But that is really just a way of reducing the monetary base to make the price level determinate. It’s like the tail wagging the dog. It looks like it is changes in interest rates that keep prices on track, but it is actually changes in the monetary base.

    2. Consider a policy of doubling the monetary base. We know that that policy will double the price level. What will happen to the fed funds rate? We don’t have any idea. We don’t even know whether a doubling of the monetary base would cause the fed funds rate to go up or down. Now suppose I asked you to double the price level via changes in the fed funds rate. Neither you or I have any idea what fed funds rate would double the price level, and I doubt anyone else does either.

    3. Consider monetary policy in a world with base money but no banks or interest rates. It is possible to imagine policy proceeding effectively through changes in the base. Now consider a world with interest rates but without money. It’s called barter, and there would not even be any price level, or such a thing as monetary policy. The base is fundamental.

    Bill, I basically agree, and if anything am even less worried than you are. I don’t even care whether we have weekly data on the monetary base. As long as the Fed can monitor inflation indicators such as yield spreads, commodity prices, etc, they can avoid overshooting. I don’t know of any inflation in history that was caused by overshooting when the central bank knew what it was doing. Of course central banks have only known what they were doing since 1982, so my “history” is pretty short.

  46. Gravatar of Jon Jon
    15. April 2009 at 19:56

    Scott and I agree well enough that I don’t see a point to nitpicking.

    JKH: thanks for taking the time to type up a nice exposition. I agree with much of what you say–and I think its consistent with Scott’s position too. Our differences lie in your penultimate paragraph. I agree that regulation of short-rates do have direct influence (such as by pegging one end of the yield curve); I’m just not convinced that this mechanism dominates the effects that arise from quantity changes.

  47. Gravatar of TheMoneyIllusion » Who is impersonating Robert Hall? TheMoneyIllusion » Who is impersonating Robert Hall?
    3. February 2013 at 19:24

    […] I started blogging Robert Hall was one of my heroes.  In this post I quote Hall discussing negative IOR and pointing out that there is no basis for the claim that […]

Leave a Reply