A rose by any other name . . .
One of the frustrations with blogging is that people focus too much on the names we attached the concepts and not enough on the underlying concepts themselves. A few months back I had to fight a long battle to convince others that the Obama administration had raised the top income tax rate to 43.4%. The administration and its echo chamber in the liberal blogosphere were engaged in the “big lie” that they had merely returned income tax rates to the level of the Clinton Administration. They excluded the new 3.8% income tax to finance healthcare, because it was called a “payroll tax”. Of course a payroll tax only applies to wages, and this tax applied to both wage and capital income. So it is effectively an income tax whatever the administration chooses to call it.
Two more recent examples popped up in the comment section. Some commenters insist that the Federal Reserve is “monetizing the debt” because they are exchanging interest-bearing reserves for interest-bearing Treasury securities. Yes, it’s true that interest-bearing reserves are considered money. However the term “printing money to monetize the debt” has a very specific meaning. It refers to the exchange of high-powered money for interest-bearing debt. It’s also true that prior to 2008 bank reserves were high-powered money. But obviously that was no longer true after the Fed started paying interest on reserves. High powered money is “high-powered” precisely because it doesn’t earn interest, and hence is subject to the hot potato effect, at least in the long run.
Why is it so important to be clear on terminology? Consider all the statements that; “previous cases where governments printed large amounts of money to buy back debt have usually resulted in hyperinflation.” As far as I know, however, there is never been a case where the exchange of interest-bearing reserves for government bonds has led to hyperinflation. What would be the point? (If there is please correct me on this.)
Another recent example occurred when someone questioned my assertion that FDIC bailouts are financed by taxpayers. Again, I think the confusion relates to terminology. The FDIC tax on banks is called a “fee”, if I’m not mistaken. However it is effectively a tax on banking services, and therefore the cost of this “fee” is clearly born by people use the banking system (and perhaps partly by banks) just as any other tax on goods and services is assumed to be born mostly by consumers, but perhaps partly by suppliers as well. This should not even be controversial. The fact that it is testifies to the power of words.
Tags:
3. October 2013 at 07:03
“High powered money is “high-powered” precisely because it doesn’t earn interest, and hence is subject to the hot potato effect, at least in the long run.”
Prior to 2008 high powered money that was in excess of required reserves could earn the fed funds rate. If any bank had excess reserves they could lend it out in the fed funds market. Required reserves, similar to now, earned nothing.
3. October 2013 at 07:15
Whether it’s buyers or sellers who bear the cost of the tax will depend on the slope of their demand and supply curves. Draw a supply/demand graph, add a tax, and it’s easy to measure who loses the most surplus.
3. October 2013 at 07:43
Scott, not disagreeing with you, but perhaps you could clarify a bit: from the following you wrote:
“Positive interest rates: The base becomes a hot potato, just as in the previous examples. IOR changes that slightly, but less than you’d think. Peter Ireland showed that the quantity theory still applies in the long run, even with IOR.”
Which was your case 5.a. here:
http://www.themoneyillusion.com/?p=23314
I had the impression that IOR really didn’t make much difference (due to the word “slightly”). Now that’s a case with positive rates, granted. So is that the difference? IOR has more than a “slight” effect because we essentially have zero rates?
3. October 2013 at 07:45
Mark A. Sadowski: your services are requested here!
“”previous cases where governments printed large amounts of money to buy back debt have usually resulted in hyperinflation.” As far as I know, however, there is never been a case where the exchange of interest-bearing reserves for government bonds has led to hyperinflation. What would be the point? (If there is please correct me on this.)”
Examples? No Examples? How do you come down on this?
3. October 2013 at 07:50
Only someone with a heart of stone wouldn’t laugh at this tale of a boy, his dog, and his health insurance;
http://delong.typepad.com/sdj/2013/10/so-i-went-on-the-federal-government-health-exchange-marketplace-this-morning-to-see-what-i-could-see.html
3. October 2013 at 07:57
“…the term “printing money to monetize the debt” has a very specific meaning. It refers to the exchange of high-powered money for interest-bearing debt.”
No it doesn’t. Look the term up in a pre-QE textbook (I used Begg, Fischer and Dornbusch) and you will find that it refers to government borrowing directly from the central bank – aka “monetary financing” in Europe. The essence is more that the exchange is off-market, with the central bank typically not being given marketable assets to the same value as the advance to government, than that zero interest money is involved. This depletes the central bank’s net worth and therefore erodes its independence.
Fed coupon passes used to involve the exchange of zero interest reserves for government bonds, but were not generally referred to as printing money. At present, the Fed is not “printing money”, although the distinction becomes a bit semantic when the Fed is such a dominant buyer of treasuries and buys very recently issued bonds. In the hypothetical case that the Fed credited even an interest-bearing Treasury account without receiving marketable securities of matching value, the term “printing money” would be appropriate in its traditional sense.
3. October 2013 at 08:02
So if the Federal Reserve hadn’t been paying interest on reserves since 2008, you would have no problem with people saying things like “the Fed’s printing money”? I think you could still dispute that because the U.S. Bureau of Engraving and Printing is actually in charge of printing money. Doesn’t the Fed always either buy or sell Treasuries and therefore either print or unprint money?
In the past 100 years, have countries that have experienced hyperinflation actually bothered to go through the act of exchanging money for government debt or does the government just sort of go to the mint with a wheelbarrow (probably armored cars actually) and take what they need?
3. October 2013 at 08:09
Ed, I think the Fed does pay interest on required reserves as well as excess reserves:
http://www.federalreserve.gov/monetarypolicy/0AD345FADDDD49A8878308C9D9202BA4.htm
3. October 2013 at 08:13
Tom,
I will look into that, but the Fed doesn’t pay interest on require reserves http://www.newyorkfed.org/markets/ior_faq.html
3. October 2013 at 08:25
Ed, link towards that bottom of your link has this title:
“Board announces it will begin to pay interest on depository institutions’ required and excess reserve balances”
3. October 2013 at 08:29
Scott –
I personally find you the most difficult to understand of the economists I read. That’s an observation, not necessarily a criticism. I am sometimes missing a connecting or clarifying sentence which would help me better grasp a concept–and I am in fact interested in better understanding what you’re saying.
I’d add that I’m not a professional economist, but I am a finance professional who studied economics many years ago. The fault may be entirely mine, but pausing from time to time to clarify a concept or associated vocabulary would certainly be welcome.
3. October 2013 at 08:30
Tom,
I stand corrected. I guess they do pay interest on required reserves.
3. October 2013 at 08:32
Tom,
http://www.federalreserve.gov/monetarypolicy/20081216d.htm
3. October 2013 at 08:32
Ed, I don’t think they pay interest on vault cash though. Am I correct?
3. October 2013 at 08:33
Bill Gross today:
“Don’t run for the hills b/c of the #shutdown or the debt ceiling – Run b/c the economy is slowing by itself.”
“” PIMCO (@PIMCO) October 3, 2013 (Twitter)
3. October 2013 at 08:34
Ed, I’m afraid that is irrelevant. The bank holding the reserves does not earn any interest on those reserves. When reserves are loaned out one bank gains and the other loses. Is a zero-sum game.
Randomize, Yes, I know that. But of course that has no bearing on whether FDIC fees are a tax. That gets at the question of which taxpayers pay the tax.
Tom, that’s a good question. When you move from no interest on reserves, to a regime with positive interest on reserves, there is a one time increase in the real demand for base money. That increase is not necessarily inflationary. Once you’ve begun paying interest on reserves, changes in the monetary base are neutral in the long run, holding the IOR rate constant. Thus prices will rise after a permanent monetary injection.
There is a separate issue of whether swapping interest-bearing reserves for Treasury debt constitutes “monetizing the debt”. I say no, at least not as the phrase ‘monetizing the debt’ is usually used.
Rebeleconomist, I disagree. When people talk about monetizing the debt and hyperinflation they are talking about the sort of policy that you saw in Latin America in the 20th century, or in Germany in 1923. You need to look at how words are used in context, not just the literal meaning.
John, I would not object to people using the phrase monetizing the debt. However I would’ve still objected to the Kotlikoff column. I would’ve disagreed with this assertion that the policy was highly inflationary, and I would’ve disagreed with his assertion the Bernanke’s motive was to monetize the debt.
I believe that in most cases during hyperinflation the vast majority of the monetary injections our currency not bank reserves, but I’m not certain.
3. October 2013 at 08:35
Tom,
My first instinct was to say “of course not” but am not so sure anymore ;).
3. October 2013 at 08:36
Steven, I’m 100% amateur at all this. I found Scott easier to understand after reading his FAQ.
3. October 2013 at 08:38
Steven, sorry about the confusion. I think one problem is that I often go against the conventional wisdom, so people coming into the my blog from the outside have trouble grasping my way of looking at the world. I’ll try to be clearer in the future.
3. October 2013 at 08:39
Scott, are you saying that you’ve joined the “just a swap” crowd?
On one hand, with no hot potato, they have likely created more reserves than if they had a HP. (Now, a hot potato would have helped immensely to boost NGDP, which I thought was the goal).
IOR therefore means a bigger balance sheet than before, which now means not only larger perpetual IOR expenses, but also a huge principal risk. Let’s call it 20% losses on $3.7t at 7 year duration and ten-year bond normalization to 5%, on top of some $70 billion in IOR per, with liberal assumptions.
All these costs go to the taxpayer, whether by reduced remittances or by the now (bigger-than-fannie-or-freddie) GSE of the Fed. It’s a asset-liability mismatch. Perpetual monetary base obligations exchanged for limited-life debt.
So how is this funded? What sort of liabilities does the Fed pay for IOR? How does the Treasury re-equitize the Fed? Like it did Fannie? These matters are a long way off, but are direct consequences of IOR, and may have inflationary consequences.
More specifically on the “just a swap” front, reserves can be taken away from banks and transformed directly into currency: currency can be withdrawn by the public, which reduces excess reserves at the banks. Federal debt cannot be transformed into currency directly, it needs to first go through the reserves channel.
Now, the Fed always accomodates the demand for currency, but there is now an immense stock of perpetual-life proto-currency: it just has not been demanded yet. Historically, first comes reserves, then comes currency. Zero rates are not a good incentive to keep cash in banks.
http://research.stlouisfed.org/fred2/graph/?g=n1a
I have not even gone into what the “just a swap” thesis means for your thoughts on NGDP at the ZLB.
3. October 2013 at 08:39
Tom and Ed, I believe they pay interest on required reserves but not vault cash. But I’m not certain.
3. October 2013 at 08:40
Okay, I’m confused. Where did the money come from that the Fed used to buy all the assets it owns (trillions of dollars of assets)? I thought it came from someone typing a 1 followed by a bunch of zeros into an account at the NY Fed (the modern version of printing money). If this isn’t where its coming from, then where is it coming from?
3. October 2013 at 08:40
So, now a question:
The Treasury issues a bond to the public, which buys it. The Fed then buys the bond from the public with a new accounting entry–it essentially creates the money. After this purchase, the public has no greater or lesser holdings of financial assets than before the bond was created.
The Treasury then spends the money it received from the public. The Treasury spends this covering the deficit, that is, spending it on things like defense, Social Security and Medicare. So that money actually makes it way into the economy, no?
For its part, the Fed then has, on its asset side, a bond payable by the Treasury, and an increase in equity on the liability side which came from nowhere.
If I understand correctly, as long as the Fed has an asset offsetting the increased equity, the effect of the Treasury’s spending for government services and goods is not inflationary. However, if the Fed were to offset assets against liabilities, and that bond and the associated equity were written off, then the money that the Treasury spent on services would be inflationary? Is that it? Is the difference between monetizing the deficit and not merely an accounting entry?
3. October 2013 at 08:41
Scott: currency pays 0% interest nominal. But is there anything magic about 0% nominal? Suppose that by custom, currency paid a fixed (say) 1% nominal. The demand for currency would be slightly higher than otherwise. But a world with a 3% inflation target and 1% nominal interest on currency would be like a world with a 2% inflation target and 0% interest nominal on currency.
Now when you vary the rate of interest on currency, the central bank has a second policy instrument, in addition to the quantity of currency.
By extension, something roughly similar is also true for interest on reserves.
3. October 2013 at 08:42
“Ed, I’m afraid that is irrelevant. The bank holding the reserves does not earn any interest on those reserves. When reserves are loaned out one bank gains and the other loses. Is a zero-sum game.”
The bank holding the excess reserves does earn the funds rate when they lend out the reserves. The bank that borrows the reserves (prior to ’08) wouldn’t earn anything on their required reserves. So no it wasn’t zero sum.
3. October 2013 at 08:43
You are free to disagree, Scott – but you’d be like Humpty Dumpty!
3. October 2013 at 08:45
Steven, for what its’ worth, the reserves the Fed issues are counted as liabilities not equity (yes, both liabilities and equity are in the “Cr” column, but they’re different). Same with cash they sell to banks. That’s how the accounting works (although some people, like Nick Rowe, don’t like that). Google Federal Reserve balance sheet and you can find an example.
3. October 2013 at 08:47
jknarr, I don’t understand your “just a swap” comment. In any case, payment of interest on reserves may expose the Fed to greater financial risk. But that’s a trivial problem because the Fed is part of the federal government. Losses to the Fed are exactly equal to gains to the federal government when bond prices fall.
Yes, reserves can be viewed as proto-currency. However the Fed can easily prevent the sort of spillover effects that would lead to high inflation.
3. October 2013 at 08:49
Why does it matter if the Fed loses money on the assets it purchases. The money came from nowhere so the losses go the same place.
I thought the whole point of printing money is to make it less valuable so that people will be willing to trade it for goods and services instead of hoarding it. Well, if money becomes less valuable (i.e., higher inflation), then it is plainly obvious that the value of bonds will fall. So, if the Fed is going to buy bonds, then the only way it will succeed is to have the value of the bonds drop.
3. October 2013 at 08:52
Steven, I can’t post a link, but if you Google
Tom Brown banking example #11
You’ll find where I put up simple set of balance sheets you might find interesting (it speaks to your comment above a little bit). I’m assuming like you do that the Tsy spends every $ it gets.
3. October 2013 at 08:52
Why are marginal tax rates always measured without reference to marginal utility? A marginal tax rate is a dimensionless number that has no meaning without such a reference. Is 5% high or low? Remember, it’s a marginal tax rate. That means 5% is low if it kicks in at a billion dollars, but high if it kicks in even if your business shows a loss. Heck, 50% at $200k would be considered low in the middle of the last century, but would be considered high today (recall the discussions in 2008-9 of whether $250k was “rich”). People default to the references of 0% and 100% but that is rather arbitrary, right?
If I take the desired precautionary savings rate as a measure of marginal utility such that someone with $1M in income has about half the utility of their last dollar as someone making $30k, then a 25% marginal tax rate on the low end and a 50% marginal tax rate on the high end both end up taxing the same amount of marginal utility (1.0*$1*0.25 = 0.25 and 0.5*$1*0.5 = 0.25).
[And that’s a conservative measure of marginal utility! Naively — i.e. at the starting point of economic analysis — one would expect someone to value their millionth dollar per year at one millionth of their first dollar per year.]
3. October 2013 at 08:59
jknarr, re: “just a swap” … I noticed something similar, but that happened yesterday, and it’s an “merely exchanging” not a “just a swap” 😀
“The Fed is merely exchanging interest-bearing reserves, which are liability of the federal government, for Treasury securities, which are different liability of the federal government.”
http://www.themoneyillusion.com/?p=23933
Sorry Scott… I think I see what you’re getting at. “just a swap” is annoying when used to imply QE does nothing… I get it. However, “merely an exchange” is useful for battling hyper-inflationists. I get that too. 😀
Or do I actually not get any of that and am just embarrassing myself )c:
3. October 2013 at 09:02
Scott N, That’s right, but I wouldn’t call that “printing money”. The phrase printing money has traditionally implied the production of non-interest-bearing outside money.
Steven, No, the issue is not the appearance of the Fed’s balance sheet. It’s the supply and demand for the monetary base. The demand for base money is affected by both the interest rate on reserves and the opportunity cost of holding cash, which is the market interest rate. You get inflation when the supply of base money rises faster than the real demand for base money.
Nick, I agree. The important factor is the difference between the interest rate on base money and the interest rate on the government bonds that are purchased with base money. I tried to avoid making the post excessively complicated by talking about that distinction because it gets into the question of how do you measure the expected yield on a long-term bonds over the next day. Right now the interest rate on reserves is higher than the interest rate on short-term government debt, and probably higher than the interest rate on long-term government debt if measured properly, accounting for the term structure of interest rates.
Ed, No, once reserves are lent out the bank is no longer holding those reserves.
Rebeleconomist, I am the one using these terms in the conventional way. I doubt you could find anyone talking about how “printing” interest-bearing reserves can lead to hyperinflation, at least prior to 2008. If you can find it in the economics literature please let me know.
You probably don’t know this, but there is a long literature on the concept of printing money to monetize the debt and how that can lead to hyperinflation. In that literature 100% of the cases refer to the production of high-powered money, not interest-bearing reserves. You can call it whatever you want, but it doesn’t help Kotlikoff’s argument at all.
3. October 2013 at 09:09
Tom, Now I see what you and Jknarr are getting at. I actually have been consistent here if you read me closely. I’ve always argued that QE as a very small impact on inflation. Not large, but not zero either. That is still my view.
The phrase “just a swap” referred to the fact that the fiscal implications are negligible. So in that sense you weren’t really financing the budget deficit by printing money. The effect on inflation depends on all sorts of other factors, such as whether the currency injections are viewed as permanent, and the degree to which bank reserves and Treasury securities are close substitutes. Markets suggest that QE has a small positive effect on expected inflation, but nothing dramatic.
3. October 2013 at 09:11
Steven Kopits,
Broadly right — look here under statistical tables for “factors supplying” and “absorbing” reserve funds.
http://www.federalreserve.gov/publications/annual-report/files/2012-annual-report.pdf
But the Fed does not buy assets from the public, they buy assets from the primary dealers in exchange for reserves (special bank money-capital). Look up the frequent “POMO”s.
http://www.newyorkfed.org/markets/OMO_transaction_data.html
On one hand, the public has money, they hand this cash to the Treasury, get a note back, the Treasury re-spends the money: a bit more NGDP, a lot more debt, same amount of base money.
But when the Fed buys a note, the primary dealer first buys it at Treasury auction (or from the public), transfers demand deposits to the Treasury (or public: it all ends up being spent on EBT cards, missiles or dinners anyway), and then sells the bond to the Fed. The Fed creates reserves to buy it from the bank: so more demand deposits, more reserves. Note that the Fed could buy bonds for currency (if it bought a bond from a guy at the corner of wall and broad), but it does not.
The primary dealers, treasury, and Fed all operate at market prices for bonds. Market prices for bonds are set by the broad level of NGDP activity in the economy. The relative scarcity of Fed liabilities (base money) is the root of value for all money throughout the economy: all other derivations of money (demand deposits) are simply promises to deliver this base money on demand.
3. October 2013 at 09:14
Scott,
I will try one last time.
After an open market operation, one bank has excess reserves and another bank doesn’t meet its reserve requirements. The bank with the excess reserves lends its excess to the bank with the deficit. The original bank with the excess reserves earns the funds rate. The bank with the deficit earns nothing.
3. October 2013 at 09:25
Scott,
Larger point is that even prior to ’08, a bank could earn the fed targeted funds rate on any excess reserve balances they had.
3. October 2013 at 09:28
Steven & jknarr, I found this useful (it’s a cut & paste from a Philippe comment):
This recent paper details who ultimately sells assets to the Fed as part of QE:
“Overall, our results suggest that the Federal Reserve is ultimately interacting with only a handful of investor types. Households (the group that includes hedge funds), broker-dealers, and insurance companies appear to be the largest sellers of Treasury securities when the Federal Reserve buys these securities. Households, investment companies, and to a lesser extent, pension funds, are the largest sellers of MBS when the Federal Reserve buys. With both the Federal Reserve’s Treasury and MBS purchases, our results suggest that households are the largest, ultimate seller…”
“Analyzing Federal Reserve Asset Purchases: From whom does the Fed buy?” (2013)
http://www.federalreserve.gov/pubs/feds/2013/201332/201332pap.pdf
3. October 2013 at 09:28
I refer simply to “printing money”. The government may or may not use that money to “monetise (its existing) debt” by buying it back.
You won’t find “printing interest-bearing reserves” in the pre-QE literature because central banks generally did not pay interest on reserves then, but as Nick Rowe says, interest is just a property of base money, and if short-term market interest rates are 5%, there is not a lot of difference in hotness between zero-interest and 25bp-interest reserves.
Although the Fed is not (yet) printing money, Kotlikoff has a point because when central bank purchases are so big that they are significantly distorting market prices, it is likely that in the long run the Fed will lose money, meaning that to some extent present government spending is effectively spending Fed capital. QE is not an off-market transaction, but the government and the Fed are most of the market! And this does increase the threat of inflation because the Fed may not be able to reabsorb surplus base money if it runs out of assets and the government will not grant it more.
3. October 2013 at 09:29
… and here’s a follow up comment from Mark A. Sadowski:
http://worthwhile.typepad.com/worthwhile_canadian_initi/2013/08/how-can-you-get-an-economy-into-a-liquidity-trap.html?cid=6a00d83451688169e2019aff5f2906970b#comment-6a00d83451688169e2019aff5f2906970b
3. October 2013 at 09:31
Jason,
Marginal utility is really the entire justification of having progressive taxation in the first place. The fact that we tend to believe that each additional dollar means more to someone making 25k a year than 250k a year is the reason we have progressive tax rates.
The reason we talk about marginal tax rates without specifically referencing marginal utility is that marginal utility is already assumed. The problem then becomes the incentive effects of high marginal tax rates.
I would also add that when talking about marginal utility people often forget that some people value money more than others. That is to say, if an artist and a banker were both making 50k (this is a hypothetical world), the next dollar would probably mean more to the banker because money is likely a larger factor in his decisions than the artist’s.
In other words, when looking at tax policy from a utilitarian perspective, we have to consider: marginal utility and economic efficiency. When it comes to marginal utility, we have to remember that not everyone values each additional dollar the same at any given income level.
Once you look at the evidence for long term incentive effects of marginal tax rates and use a more realistic view of utility, the Krugman/Picketty arguments that 70-80% tax rates are optimal seem quite absurd, as they should.
3. October 2013 at 09:32
OK, I understand this. But the question I have is, WHY did the Fed launch QE rather than simply stopping payments of IOR? Or at least you would think they would have stopped paying interest on reserves and test the effect of that before proceeding with the “dangerous” QE program.
3. October 2013 at 09:38
Is it the payment of interest on reserves that matters here, or is it that, with the payment of interest, reserves and money are close substitutes? Suppose money and government bonds were perfect substitutes and the interest rate is zero on both. Buying bonds (say those held by banks as excess reserves) with money would have no effect on, well, anything. Only the new issue of high power currency, by expanding bonds plus money supply, would have an effect (as would issuing new debt). Am I correct on this? I am not a macro economist, so corrections that improve my understanding are welcomed.
3. October 2013 at 09:47
Tom Brown, yes I’ve seen that. It’s all correlation and inference, and very misleading as a result. The FoF data is broadly useful, but has nothing to do with who is shaking hands with the new york fed. That, and the household segment is notoriously “grab-bag-residual”.
The first indicator of who the Fed is buying from is the composition of their payment liabilities: swaps, reserves, currency: it needs to be in acceptable form (if the Fed creates demand deposit liabilities for households and hedge funds and insurance companies, its not listed on their balance sheet). We can also see a mirrored “cash assets” in the H8 report for the commercial banks: the Fed says it creates reserve liabilities, and the banks report matching cash-equivalent asset holdings.
So we can see a bit of who is getting reserves – in exchange for assets. The Fed has been providing reserves (in exchange for what, again?) almost exclusively to foreign banks in QE2 and QE3. QE1 was the only domestic operation, with large increases in cash equivalent assets / reserves at US domestic banks. The breakdown is domestic large/small bank and foreign bank.
http://www.federalreserve.gov/releases/h8/current/default.htm
H.8; Page 20
Assets and Liabilities of Foreign-Related Institutions in the United States
3. October 2013 at 10:00
“The FDIC tax on banks is called a ‘fee’ . . . . However it is effectively a tax on banking services, and therefore the cost of this ‘fee’ is clearly born by people [who] use the banking system (and perhaps partly by banks) . . . .” However, in return for this payment the banks and their customers get a special service: the government guarantee of deposits. I assume that firms have the option of operating as banks outside the FDIC safety net, in which case the FDIC fee does, indeed, seem to be a payment for a service””i.e., a genuine fee. (The scarcity of non-FDIC banks suggests that the fee may even be too low.)
3. October 2013 at 10:08
Scott,
Is it LNGDP targeting that president Fisher is advocating in his latest speech?
“My point is simply to highlight the longer-term consequences of what might appear to be smallish, shorter-term deviations from the norm. Business operators plan capital expenditure and payrolls not in one- or two- or even three-year increments; they plan and budget over longer-term horizons. The nominal stability that people need if they are going to negotiate multiyear contracts is a multiyear nominal stability. A policy that “lets bygones be bygones” from year to year may not achieve this kind of stability, especially when policy options can become constrained, in the short term, by the zero bound. A policy that takes a longer-term perspective and is properly communicated and executed””so as to instill confidence that monetary policy will hew to a 2 percent inflation target rather than fixate on the run-rate of the past four quarters or the outlook for the next four””may better supply the longer-term comfort that households and businesses need to plan and budget. ”
http://www.dallasfed.org/news/speeches/fisher/2013/fs131003.cfm
3. October 2013 at 10:09
jknarr, if you’re saying that the Fed doesn’t create bank deposits directly (for non-banks) and must use Fed deposit holders (i.e. banks) as intermediaries, whose balance sheets also grow as a result (w/ reserves in the assets column and deposits in the liabilities column), there’s no argument from me on that! In fact I demonstrate that in a simple way as “Case 2” in my “Banking Example #4.1” which I doubt I can link to… but you can google it (throw “Tom Brown” in to get there). The interesting thing to me was that banks end up holding a relatively small percentage of Tsy debt (if I’m interpreting that right)… i.e. they do all the buying and selling, but mostly act as intermediaries to the broader market (and yes, I’m conflating PDs w/ banks here). That’s what I got out of it anyway. Is that basically correct?
3. October 2013 at 10:10
Ed, You said;
“Larger point is that even prior to ’08, a bank could earn the fed targeted funds rate on any excess reserve balances they had.”
Sorry, but this is flat out wrong, and I think all economists and bankers would agree with me. When you lend out ERs you no longer “have” them. You are earning interest on a loan you have made, that’s totally different from earning interest on reserves.
Rebeleconomist, You said;
“as Nick Rowe says, interest is just a property of base money, and if short-term market interest rates are 5%, there is not a lot of difference in hotness between zero-interest and 25bp-interest reserves.”
Yes, and as I explained that has no bearing on this post. Interest rates are currently 0% on T-bills. The IOR is higher.
As for whether the Fed is likely to lose money, if you’re so smart, how come you aren’t rich? And if you are rich, why are you wasting time in my comment section?
dbeach, Great question. And here’s an even better question, why did they start IOR in October 2008 just as nominal GDP was collapsing and asset prices were plunging? Even the Fed concedes that its 2008 decision was contractionary. So why do it?
Walter, Yes that’s basically correct. If cash and bonds are perfect substitutes, than bonds are essentially cash. So you’d be swapping cash for cash g
3. October 2013 at 10:19
Philo, This link says they are all covered, but I am not sure.
http://www.ehow.com/about_4600324_banks-that-not-fdic-insured.html
Thiago, That’s the logic behind NGDPLT, but he says he favors price level targeting.
3. October 2013 at 10:23
Finally, Scott, I can’t help thinking that “debt monetization” is just a bad concept. I’ve been influenced by: http://fofoa.blogspot.com/2010/09/just-another-hyperinflation-post-part-3.html Not as systematic as you’d like, ignore the gold if you wish, but conceptually useful.
Suggests that large-scale inflation is mostly a change in the “store of value” away from financial assets and into real assets. When the (now-immense) wealth/value of financial assets is threatened, it is sold, and real assets bought, and so base money is demanded and produced in between (as we have seen since 2008) to liquify the transaction demand. Money is never ever a store of value, and it is only a waystation, so “monetization” is a useless concept — there is only medium of exchange demand for money, and value is kept in assets, whether real- or financial-.
Financial asset liquidation sales creates a huge amount of demand for the medium of exchange. But because the two (exchange/account) are the same, unfortunately, this also creates a surplus of the medium of account. The hyperinflationary physical cash only shows up later as a rump of exchange demand — there is never enough on hand to keep up with the rapid real asset price acceleration set by the deteriorating expected medium of account.
So widespread financial asset liquidation demands medium of exchange base money into existence, which dilutes the medium of account, which leads to hyperinflation and a re-balancing of the equilibrium between financial asset promises and real-world productivity and assets.
IOR is meaningless in this hyperinflationary context. They could raise IOR and not provide the base money demanded by a liquidation, but this would only mean more panic selling and illiquidity: nobody to buy, and would make the problem worse. They know this, and so will inevitably provide base money when liquidation pressures mount. It’s not “monetization”, just providing monetary liquidity on demand: which was the entire point of the Fed, anyway.
3. October 2013 at 10:30
Scott,
What was the purpose of the fed funds market prior to ’08?
3. October 2013 at 10:32
Wiki isn’t necessarily the final word, but I believe this to be accurate:
“In the United States, federal funds are overnight borrowings between banks and other entities to maintain their bank reserves at the Federal Reserve. Banks keep reserves at Federal Reserve Banks to meet their reserve requirements and to clear financial transactions. Transactions in the federal funds market enable depository institutions with reserve balances in excess of reserve requirements to lend reserves to institutions with reserve deficiencies. These loans are usually made for one day only, that is, “overnight”. The interest rate at which these deals are done is called the federal funds rate. Federal funds are not collateralized; like eurodollars, they are an unsecured interbank loan.”
3. October 2013 at 10:55
Ed, I think the point Scott is making is that in order to collect your Fed funds rate prior to 2008 (if you’re a bank lending out reserve to other banks) you had to part with your reserves. If you had and kept excess reserves… you got diddly for them. That’s the big difference. Earning a big fat 0% on your excess meant that you try to get rid of them for something that pays better. So if the Fed didn’t pay IOR and they did QE, the effects of QE would be stronger (in theory). But as it is now you keep your reserves to earn IOR. HPE is weaker.
3. October 2013 at 11:05
Tom,
But why would anyone keep excess reserves if they can earn anything. If there were no IOR right now, the excess reserves in the banking system would force rates to zero. The Fed doesn’t want HPE below IOR. For whatever reason the Fed prefers to keep rates slightly higher than zero. Because of frictions, the banking system cannot fully arbitrage IOR to other overnight rates like fed funds and general collateral. So now they are considering other facilities to overcome that http://www.newyorkfed.org/research/staff_reports/sr642.pdf
The point is that in the pre 2008 system, the return on excess reserves for a bank was the funds rate, and now it is IOR.
3. October 2013 at 11:10
Tom – Thanks. I am always amazed what you can find on the internet. Thanks for you interesting site.
3. October 2013 at 11:10
“However, in return for this payment the banks and their customers get a special service: the government guarantee of deposits.”
I suppose that’s why it’s called a “fee”? Oh, well, a rose is a rose.
3. October 2013 at 11:16
Well I have a cynical reason to explain the Fed starting the payment of interest on reserves in October 2008, but I was hoping there was a better one. I always have believed that IOR was a backdoor way of recapitalizing the banks by just handing them free money at a time when many of the largest institutions were on the verge of collapse. Even if that’s the case, though, the large banks do not have solvency issues now, so I don’t see why they wouldn’t have stopped paying IOR instead of or at least prior to launching the QE program.
3. October 2013 at 11:19
Scott –
OK. So let’s see if we’re in agreement. The Fed is creating money, which is cycled through the Treasury and is spent on goods and services. So the money really does find its way to the real economy. Yes?
Now, the question is then the demand side. How does this economy interpret this new supply of money? It could
i. …show up as decreased interest rates and support investment in fixed goods; or
ii. …increase the production of goods and services if there is slack in the economy; or
iii. …show up as inflation if people aren’t interested in what the unemployed have to offer; the sum is greater than the slack in the economy; or people aren’t interested in investing in fixed assets.
So, depending on the circumstances, the new money could support economic activity, or could lead to an asset bubble, or could lead to price and wage inflation. Is that right?
3. October 2013 at 11:31
Ed, yes I know that ER > 0 (for the banks in aggregate) implies FFR = IOR.
I’m just trying to clear up what seemed to me like a miscommunication between you two. You’re thinking in terms of individual banks and Scott is thinking in terms of aggregates here… I see what you’re both saying (I think).
Your point is that prior to 2008 there was an FFR… and to the extent that there was a mismatch of deposits between individual banks they paid that rate… or they paid it in interacting with the gov or the Fed… etc.
Scott’s point is that in aggregate, prior to 2008, ER = 0 (essentially) and IOR = 0. If you’d kept IOR = 0 and increased aggregate ER > 0, now there’s an HPE. There was no HPE prior to 2008 in aggregate because ER = 0. Well actually there was an HPE but ER was close enough to 0 that only small changes in reserve levels by the Fed could be used to adjust the FFR.
My point is I don’t think Scott misunderstands that prior to 2008 a bank could lend it’s ER to other banks and earn the FFR for it. He’s saying that it’s not valid to bring that up to say that IOR doesn’t matter now.
Shoot… I’m having a hard time expressing this.
Basically post-2008 (ER > 0), IOR > 0 implies HPE weakened. This has nothing to do with individual banks lending to each other prior to 2008 (ER = 0) when IOR = 0, and FFR > 0.
3. October 2013 at 11:44
Tom,
I think you are expressing things rather well. There is a key point that I would like to clarify.
There was a minimum rate (fed funds) pre 2008 as well and there was HPE to that rate. To the degree fed funds on a given day went above the Fed’s target, the NY Fed would perform open market operations and increase reserves. Those new reserves would then HPE until the banks without adequate reserves got them and stopped bidding for funds above target.
So I would disagree that HPE weakened. It is all about the overnight rate the Fed is targeting and the tools that would help them achieve that.
3. October 2013 at 11:49
Ed, but when IOR > the rate on Tsy debt… how does that not weaken the HPE?
3. October 2013 at 11:57
Tom,
As I said earlier there are frictions that keep banks from perfectly arbing IOR with other overnight rates (balance sheet limitations, FDIC fees, borrowing USTs when special, etc).
I suggest you read sections 1,2 and 5 of that NY Fed paper on how they are trying to overcome the slippage between overnight rates. The most important spread is between IOR and GC. GC is simply the overnight rate on non-special, collateralized US Treasury loans. These new facilities, especially RRP, should help narrow that range.
So yes IOR isn’t as significant or nefarious as many make it out to be It is simply a tool to put a floor under overnight rates when banking system has excess reserves.
3. October 2013 at 13:47
Ed, I can certainly see a motivation for IOR > 0. I don’t think of it as “nefarious.” Now is it somewhat self-defeating in terms of positive benefit to QE? I can see an argument there too.
3. October 2013 at 16:02
jknarr, You completely lost me. If there was more demand for base money where does the hyperinflation come from?
Ed, I agree with that quotation, but that’s not what you were saying. Tom is right. You are only paid interest from the Fed on reserves you hold, not those you lend out. Think in terms of the consolidated banking system. With IOR the system as a whole is paid interest from the Fed. With no IOR, there is no net payment of interest from the Fed to the banking system. In the Fed funds market one bank pays interest, the other bank receives interest. IOR creates more demand for reserves. Interbank loans do not create a net demand for reserves. One bank has more demand, the other is less demand.
Vivian, Yes, and from the gas tax you get nice new roads. So it’s not a tax?
Next time I’m at the station I’ll tell the gas station attendant that I prefer not to pay the tax, after all I don’t mind potholes. And the gas tax is just a “fee”.
dbeach, I honestly don’t think that was the motive. Ben Bernanke was pretty angry at the banks at the time. I doubt he was looking to do them any special favors. And the rate wasn’t really high enough to have much of an impact in terms of saving the banking system. The Fed had asked Congress for permission to pay interest on reserves long before the crisis. They had other motives.
Steven, You said,
“OK. So let’s see if we’re in agreement. The Fed is creating money, which is cycled through the Treasury and is spent on goods and services. So the money really does find its way to the real economy. Yes?”
No. It’s sitting in the banks as excess reserves.
3. October 2013 at 16:07
“Scott: currency pays 0% interest nominal. But is there anything magic about 0% nominal? Suppose that by custom, currency paid a fixed (say) 1% nominal. The demand for currency would be slightly higher than otherwise. But a world with a 3% inflation target and 1% nominal interest on currency would be like a world with a 2% inflation target and 0% interest nominal on currency.”
This is 2% (productivity gain) deflation, 3% RDGP with a NGDPLT of 1%, right?
I need trying to noodle what happens when we have reach peak leveage.
Basically, a different take on Greenspan’s “weight” of the economy.
If we share driverless cars, and Harvard edu is free online, and atomic consumption decreases and digital consumption increases.
And there is no scarcity in the digital.
What do we do if there’s simply nothing left to lever? If there aren’t enough things to dice and splice, and the money is freakig out bc it has to go somewhere?
If we have truly crested the high water mark of things to lever…
Doesn’t the NGDP level target go down? while we lap up as much RDGP as possible (shrinking govt) since that lets us have more deflation?
Can anyone explain what happens then?
Really asking
3. October 2013 at 16:09
And… as long as we are saying that the 2008 IOR is a fundamental and profound change to the monetary base, we should also keep in mind that bank deposits are now also fundamentally and profoundly changed with the 2005 bankruptcy act and 2009 frank-dodd, i.e. subordinated to derivatives.
Many countries, including your FDIC, are now also moving toward bail-in regimes. Bank deposits are effectively unsecured subordinated liabilities: the world of fantastic e-currency at work.
http://www.fdic.gov/about/srac/2012/gsifi.pdf
http://www.financialsense.com/print/contributors/john-butler/someone-has-got-to-pay-will-it-be-you
So, the incentive is now, quite bluntly, to use derivatives for private gain, and and seize deposits to socialize losses.
http://www.stanfordlawreview.org/sites/default/files/articles/Roe-63-Stan-L-Rev-539.pdf
No wonder they are building up currency for ATMs: the real base money.
http://www.ft.com/intl/cms/s/0/11d59fa8-2c38-11e3-acf4-00144feab7de.html
3. October 2013 at 16:38
ssumner, as far as contributing to hyperinflation, excess reserves will be about as effective as short term government debt. When people stop rolling over bonds the central bank will be the only buyer, and buying with new money. When people stop rolling over bonds the excess reserves will also come out into the real economy (as cash if need be). Both of these will add trillions to M1. Both of these will contribute to hyperinflation.
http://howfiatdies.blogspot.com/2013/02/excess-reserves-is-like-government-debt.html
http://howfiatdies.blogspot.com/2013/09/usa-hyperinflation-risk.html
3. October 2013 at 16:54
Scott would you disagree with this:
Larger point is that even prior to ’08, a bank could earn the fed targeted funds rate on any excess reserve balances they had BY LENDING THEM OUT.
(I thought that was obvious).
Please read the NY Fed paper. And if you have time read the Martin, McAndrews, and Skeie paper. You are wrong on IOR. With excess reserves rates would go to zero. It is a way for the Fed to have control on overnight interest rates. And if the Fed could solve the paper currency issue, it could be a way for the Fed to charge banks if they hold excess reserves (by making IOR negative).
Tom,
It is precisely my point that IOR doesn’t in any way reduce the effectiveness of QE. QE, to the degree it is effective, doesn’t depend on the banking sector. On that, I believe Scott would agree with me. Ignore the banking sector when analyzing QE.
3. October 2013 at 16:57
Scott, I’m suggesting that financial asset liquidation can call into existence a huge amount of MOE base money, kind of like what happened with QE1, and thus cause a across-the-board MOA deterioration, i.e. inflation.
It’s a question of demand: either governments can demand huge volumes of base money to fund operations (essentially MOE), or financial asset (private) investors can demand huge volumes of base money for MOE.
Any entity that can demand and receive base money in size can cause hyperinflation: both the markets and the governments.
In either case (when one unit is used for both MOE and MOA) the demand-led MOE creation of base money leads the MOA value lower, sparking inflation.
3. October 2013 at 17:05
jknarr, You say that markets could cause hyperinflation. Do you know of any historical case of what you are talking about? Are you sure the government deficit was not being funded by the central bank in the case you are thinking of?
3. October 2013 at 17:17
Yeah, I certainly concede that they may have had other motives for IOR then and surely must now, but what are they?
3. October 2013 at 17:34
Thought-provoking blogging.
If IOER is so powerfully neutralizing perhaps a much more visible plank of Market Monetarism should be the reduction or elimination of IOER. Sumner has mentioned this idea but not really pushed it or highlighted it.
3. October 2013 at 17:41
Morgan, Yes, it is the gap that matters.
Ed, I’m running out of ways to make my point, maybe someone else can help.
Vince, No, the idea that hyperinflation is coming is far-fetched. Look at 30 year bond yields!
Jknarr, No, when the base increases due to higher demand for base money, it is not inflationary.
dbeach, Initially to keep interest rates from falling below target, as they were worried about inflation (I kid you not–read the minutes) Later they kept doing it to keep the MMMF industry afloat.
3. October 2013 at 17:44
Ben, I’ve thought of that, but here’s the problem. If the Fed eliminates IOER they are likely to do some offsetting action to neutralize the effect. People will then assume (wrongly) that it had no effect. Back to monetary offset. The problem is the target, not the tactics.
3. October 2013 at 17:56
“If the Fed eliminates IOER they are likely to do some offsetting action to neutralize the effect. People will then assume (wrongly) that it had no effect.”
Geesh! This is incorrect. The only effect eliminating IOR would have is slightly lower overnight rates. Excess reserves and bank lending would remain the same.
3. October 2013 at 18:03
Vince, debt is both someone’s asset and someone’s liability. Most hyperinflation discussions look at the size of the liability side and spending, I’m considering the size of the asset side and selling.
Pick any instance, they all have the public attempting to preserve wealth by exiting fixed incomes (bonds or wages), demanding base money and buying real goods as fast as they can(warm beer inflation), because they won’t have enough currency tomorrow: this is the cause of hyperinflation.
It’s just not simply that the government spends out of base money: its that expected MOA for income is debauched.
The ensuing demand for real goods is proportional to the value of financial wealth and income to be protected. MOE expands to accomodate this transaction flow from, say bonds and wages, into real goods: they print the cash so that the steelworker can buy all his beer at the beginning of the night, and so avoid the hourly price increases, and this cycle turns vicious between MOE and MOA.
So, if, say, total debt was 3.5x NGDP, then debt value might need to fall, and/or NGDP rise to reestablish a traditional 1.5x level (note: reality). That is, financial promises need to be ultimately backed by real productivity: it’s a part of deleveraging, and can arrive slowly, or all at once.
Still, consider that a 2% 10 year US Treasury bond denominated in an unbacked-by-gold currency might be sold faster than you imagine (at least we’ve never seen such a combination before).
http://upload.wikimedia.org/wikipedia/en/8/88/The_Hanke_Krus_Hyperinflation_Table.pdf
Nice beer color from Weimar.
http://thirdparadigm.org/doc/45060880-When-Money-Dies.pdf
3. October 2013 at 18:36
The money supply and volume of spending rises because of the Fed monetizing government debt.
Monetizing debt has a very specific meaning, yes, but it’s not what you say it is. It means the government creating new money to buy its own previously issued debt. Whether or not interest is paid on reserves and all the rest is irrelevant.
3. October 2013 at 18:53
ssumner, I think there are some flaws in your idea of using 30 year bonds tells you if there is a risk of hyperinflation. The flaw is that the central bank is spending huge amounts of new money to buy up long term bonds and prop up the price. This is not a free market price you are looking at. It is a centrally planned price.
The other flaw is that the hyperinflation seems to start with a bond panic. Before the bond panic starts there may be no clue that a bond panic is coming.
Over and over again bond holders lose big when hyperinflation starts. So if you are trying to claim that an efficient market theory means bond holders predict hyperinflation ahead of time, I don’t think the historical data supports that.
3. October 2013 at 20:36
Scott-
Well…I see your point, but we are in danger right now of the public, and even (evidently) a large fraction of the economics profession, concluding that QE is ineffective, even contractionary.
I agree that MM and that NGDP targeting are best, but in current context I would settle for a more-aggressive Fed…and robust real growth…
You got guys like John Cochrane gleefully chortling that QE is a failure and maybe even worse…
If Cochrane’s perception prevails, we will never see successful Market Monetarism, as MM requires QE when you get in or near ZLB…and I am beginning to suspect that QE may have to become a permanent part of the central bankers toolkit, given the long-term direction of sovereign debt yields….
but if QE is hamstrung by IOER…people will say QE didn’t work…
Oh, this is ugly time…..
3. October 2013 at 23:42
Over the last 50 years or so the CPI level has increased roughly 7 fold while M0 has increased about 20 fold. This seems to show that your assumption that an increase in high powered money is matched by the same increase in prices. Am I missing something?
People that set prices in the economy dont monitor the monetary base in order to set prices. There is many elements in the equation. It doesnt make sense that you can narrow down prices to one variable which is the base.
4. October 2013 at 00:00
Benjamin Cole
I agree the fed needs to print more. The best way to print and transmit money into the economy IMO is for the fed to create the mechanism to directly interact with the broader economy.
If under an inflation targeting regime when inflation is too low the fed can expand fed funds directly into accounts of public on a non debt basis. Obviously the rate of expansion is with reference to inflation.
4. October 2013 at 01:28
Scott, I comment here because I doubt some of your ideas, and since you vigorously and often cleverly defend them, I might learn something from debate with you. I soon realised that you are unlikely to take on board much from me, but some of your other commenters might. I comment less these days because your initial positions and responses to critics are so dismissive that there is less chance that I can learn from you, and because when NGDP targeting is less of a hot topic, it matters less to try to balance the discussion here.
4. October 2013 at 01:41
@dbeach, There were at least a couple of technical motives for paying IOR, for which reasons, other central banks had started paying IOR and the Fed was moving in that direction years before the financial crisis.
One is the fact that, with reserves requirements, paying no interest is a tax on the banking system, which troubled central bank economists in favour of free financial markets (remember this is pre-crisis).
The other is that paying IOR means that moneymarket interest rates tend to be more stable, which is why the BoE began to pay interest on reserves in 2006. I explain this in more detail here: http://reservedplace.blogspot.co.uk/2009/04/easing-understanding.html
(particularly paragraphs 15 to 24; note also the pre-QE definition of the term “printing money” in paragraph 41)
4. October 2013 at 03:52
Ed, You are assuming that monetary policy works through the interest rate and bank lending channel, but it doesn’t.
Vince, No, the Treasury bond market is a 100% free market. There are no price controls. And BTW, the Fed’s holdings of Treasury securities is not much different (in percentage terms) from 10 years ago, or 20 years ago.
You said;
“Before the bond panic starts there may be no clue that a bond panic is coming.”
So how do you know, if the bond market doesn’t? Are you smarter than the bond market? No offense, but no one on Earth is smarter than the bond market.
Ben, Given that the markets are proving Cochrane wrong, I think there is very little chance that his views will prevail in the long run.
Mike, You said;
“Over the last 50 years or so the CPI level has increased roughly 7 fold while M0 has increased about 20 fold. This seems to show that your assumption that an increase in high powered money is matched by the same increase in prices. Am I missing something?”
You’ve missed two big things, both related to my ceteris paribus assumption:
1. Real GDP has grown sharply.
2. Interest rates have fallen close to zero, reducing base velocity.
Obviously the RGDP growth is the bigger factor here. Any causal assertion in economics is a ceteris paribus claim.
You said;
“It doesnt make sense that you can narrow down prices to one variable which is the base.”
And show me where I did that. You need to read more carefully.
Rebeleconomist, You said;
“I comment less these days because your initial positions and responses to critics are so dismissive”
Seriously, you are accusing me of being “dismissive?” Have you ever looked in the mirror?
When commenters are polite to me, I’m polite to them. You are one of the most rude commenters.
4. October 2013 at 05:10
ssumner, “Vince, No, the Treasury bond market is a 100% free market. There are no price controls.”
The Fed spending $84 billion per month buying up bonds should at least be viewed as price supports. It is not a 100% free market.
I don’t know when the bond panic will start either.
However, all bubbles pop eventually. The Fed is artificially driving up the price of bonds and driving down interest rates. When governments try to force prices down they create shortages and when they try to force prices up they create a surplus. Eventually the production of bonds by private companies and government will increase enough to where the bubble pops. If I was able to tell the when on this kind of thing, then my yacht would not just be 26 feet long. 🙂
4. October 2013 at 05:39
ssumner, I make a big distinction between understanding the mechanics of hyperinflation and knowing when the hyperinflation will start. I understand the mechanics well enough to simulate it:
http://howfiatdies.blogspot.com/2013/03/simulating-hyperinflation.html
However, knowing when the positive feedback loop will start is knowing when a human panic will start. This is very hard if not impossible.
Understanding hyperinflation is like understanding how how fires spread that a particular forest is very dry and at risk of fire. Knowing when hyperinflation will start is like knowing when the fire will start.
4. October 2013 at 05:59
ssummers
I dont understand how withstanding ceteris paribus a growth in base will equal an equal growth on prices when most transactions dont occur in base. Broad money I agree. Most transactions I conduct are just electronic transfers through my debit card for example. I dont carry notes and coins for all my purchases. How can most of the money supply be ignored do the transactions I conduct through my debit card like many others not affect prices as well?
4. October 2013 at 06:15
Vincent, you write:
“However, all bubbles pop eventually.”
I don’t think Scott believes in bubbles, so first you’re probably going to have to convince him they exist before you can convince him that they always pop.
I realize that $84B a month is nothing to sneeze at, but what exactly is your criteria for a market to be “centrally planned” as you put it? Zero government participation? City Hall buys coffee… does that make the coffee market centrally planned and thus on the verge of being a bubble? How about the aircraft carrier market? A bubble ready to burst?
Plus I’d love to see you address this statement from Scott:
“And BTW, the Fed’s holdings of Treasury securities is not much different (in percentage terms) from 10 years ago, or 20 years ago.”
1. Do you agree that it’s true?
2. If you do agree, then how is the bond market more “centrally planned” (and thus more prone to bubbles) today than it was 10 or 20 years ago?
4. October 2013 at 06:17
I still don’t get it.
How can the Fed create money, cycle that through the Treasury, spend it on services, and still be in reserves?
I’ll have to sit through this again when I have a spare minute. I just don’t understand how the Fed can be financing a significant share of the deficit without that money ever hitting the real economy.
Why shouldn’t we then just have the Fed fund 100% of the deficit and avoiding increasing debt held by the public? There seems to be some free lunch thing here that I don’t understand.
4. October 2013 at 06:25
Vince, the level of treasury yields is almost entirely set by NGDP (and MZM, if you want to be technical). Fed buying and Treasury issuance volumes have only trivial effects on bond prices.
You can even argue that the entire reason that the Fed holds treasures is to NOT AFFECT the secondary market price of whatever asset they hold. The Fed’s power is in its creation of monetary liabilities, not its purchasing of assets.
4. October 2013 at 06:35
Vincent, if the purpose of QE central planning was to bring down long term bond yields, then weren’t QE2 and QE3 failures in this regard? Didn’t the opposite happen during those programs compared to the periods of time immediately preceding those two programs?
4. October 2013 at 06:50
Steven, take a look again at my simplified BSs for my Example #11. Again, it ignores some important stuff, like foreign holdings, social security (SS), and MBS to name a few!… but it still might help you sort this out:
I’m defining the public as the non-bank private sector, and their money as the money they control (not SS held T-bonds). And of course I’m assuming that Tsy spends every $ they bring in immediately. Then:
public’s money stock = bank deposits + cash (C) = L + B + F
L = loans
B = bank held Tsy debt
F = Fed held Tsy debt
Then if excess reserves (ER) > 0 (my case 1), we have for the banks:
reserves = F – C
but of course the banks also have other assets:
B Tsy debt
L loans to the public
To simplify my BSs, just set Ut = D = 0: that’s the case I use in the interactive spreadsheet in that post. You’re mostly interested in case 1 (ER > 0) I think.
Hope that’s of some help!
4. October 2013 at 07:04
Vincent,
Also, in Scott’s Kotlikoff post, he makes this point:
“And of course you ignore the fact that if interest rates have gone up by 2% then nominal GDP growth will go up by more than 2% because they are at the zero bound. So the Japanese government would actually be better off in that scenario.”
What is your response to that? I didn’t see that you had responded to this point.
4. October 2013 at 18:01
Tom, I don’t know if the Fed’s holdings as a percentage of the total debt are similar to periods in the past or not. I don’t see this as a factor for hyperinflation, and you know that is all I really care about. Over the last 4 years they did buy at a rate equal to a large fraction of the new debt creation. I think this may be a clue that there is risk of hyperinflation.
I doubt that the USA has ever before seen such intense monetization during peace time.
There is clearly a spectrum ranging from 0% government involvement to 100% government involvement. For long term bonds the fraction that the Fed is buying and the fraction that the Fed owns are way above 0%.
If interest rates going up by 2% insures that nominal GDP goes up by more than 2% then I think they should put interest rates up by 20%. I don’t understand that claim.
We can’t really run the experiment over without QE2 and see what the interest rates would have been that way. This is part of what makes economics so difficult. In economics you can not usually reproduce someone else’s experiment in your own lab in an afternoon.
My view is that in the short term a central bank can force interest rates down by making lots of money and buying up bonds. However, in the long term they either get more inflation than they want or higher interest rates than they want. I think the long term pain will be more than the short term gain. So if I were king the central bank would not be doing this.
4. October 2013 at 18:06
Maybe they meant that 2% inflation insures that nominal GPD goes up by more than 2%. This clearly does not hold during hyperinflation. The governments seem to always try price controls which make supply not equal demand all over the place and real GNP goes down.
4. October 2013 at 18:19
Vincent, I don’t understand this part:
“If interest rates going up by 2% insures that nominal GDP goes up by more than 2% then I think they should put interest rates up by 20%. I don’t understand that claim.”
Why do you think they “should put interest rates up by 20%” in that case? Where did you come up with that number? I’m going to ignore it for now (although I hope you respond later). I read Scott as saying essentially that you’re ignoring the possibility that NGDP will rise by more than 2% (which apparently Scott thinks is probably if interest rates do rise to 2%). Won’t this rise in NGDP tend to automatically bring in more revenue to the Japanese government to compensate for the rise in interest rates? Do you think that such a rise would be inadequate to compensate for the higher rates?
Also, do you have a counterexample for this statement from Scott:
“however, there is never been a case where the exchange of interest-bearing reserves for government bonds has led to hyperinflation.”
Isn’t that the situation in the US?
4. October 2013 at 18:21
Vincent, I didn’t see your updated comment about the 2% until after I posted: but 2% is the target inflation isn’t it? Whey would they impose price controls?
4. October 2013 at 18:47
Vince, I’m not asking for an exact prediction, but surely the 30 year bond market would be signaling a problem, wouldn’t it? That’s a long time.
And the Treasury market is much bigger than you think, the Fed doesn’t own a very large share of T-bonds.
lxdr, The neutrality of money does not in any way require that that type of money be used for transactions. The key is the supply and demand for the MOA, which is base money. Yes, money substitutes can affect things, but it doesn’t change the presumption that money is neutral in the long run.
Steven, Fed money does not “cycle through” the Treasury, it goes right into banks ERs.
4. October 2013 at 19:20
ssumner
“lxdr, The neutrality of money does not in any way require that that type of money be used for transactions. The key is the supply and demand for the MOA, which is base money. Yes, money substitutes can affect things, but it doesn’t change the presumption that money is neutral in the long run.”
Let me clarify it a little. Assuming ceteris paribus. Say over a period of time currency increases from 100 billion to 110 billion and substitutes stay the same because of ceteris paribus. Assume 90% of money is substitutes and 10% is currency. Therefore currency went up by 10% but overall money supply only went up by 1%. We arent going to have a 10% increase in prices are we? We should have a 1% increase right?
The substitutes have almost the same effect on prices as currency becuase they are used and accepted as a medium in the same way currency is by most of the economy.
5. October 2013 at 04:23
Mike, Ceteris paribus means a stable C/DD ratio, so if money is neutral all monetary aggregate rise in proportion.
5. October 2013 at 04:36
The Fed wanted to pay interest on reserves, particularly required reserves, long before the financial crisis. By doing so, they hoped to slow the movement of financial intermediation from the regulated banks to their unregulated competitors, such as money funds.
Requiring banks to hold a certain percentage of their deposits in non-interest bearing accounts at the Fed is a tax on one particular kind of financial intermediation. Bank competitors that are not subject to the tax can offer better returns to the public, and this in turn means that an increasing percentage of financial activity takes place outside of the banking system, where it is harder to monitor and control. Is it any wonder the Fed wanted to stop making this worse?
5. October 2013 at 07:07
The 10 year bond is mostly a function of NGDP and MZM, not Fed buying.
http://research.stlouisfed.org/fred2/graph/?g=n5i
So where is MZM coming from?
http://research.stlouisfed.org/fred2/graph/?g=n5j
Up until 1998, mzm had a reliable relationship with base money, both currency and reserves.
Then the Fed inexplicably shut down reserve formation from 1994-2007, and has only really returned – to the 1918 trend – with QEs since 2008.
There’s your level targeting, Scott.
http://research.stlouisfed.org/fred2/graph/?g=n5l
And you wonder why we had a banking crisis?! The Fed created the problem. Are they so stupid as to not notice? No.
5. October 2013 at 13:17
Scott,
Regarding the cycle of QE funds, this entire dialogue has been very helpful for me. 2 questions:
1) Isn’t there a good reason to distinguish between the purchase of T’s and MBS’s? A bank is likely to write more mortgages as a middle-man, earning an origination fee, knowing hat the loan will be packaged turned over. If so, the $40B/month targeting MBS would cycle through the economy and increase the base.
2) Can we not simply measure the amount of reserves sitting at the Fed and compare this to the $85B/month QE to test this hypothesis – that QE funds stagnate in reserves and are therefore base neutral?
5. October 2013 at 16:10
ssumner
“Mike, Ceteris paribus means a stable C/DD ratio, so if money is neutral all monetary aggregate rise in proportion.”
Isnt the assumption of ceturis paribus being stretched there? Assuming the private sector is going to expand deposits at same rate seems too much. Its not like the fed controls deposit creation.
6. October 2013 at 05:49
Dustin, I don’t understand what you mean by “base neutral”. Reserves are part of the base.
The Treasury has guaranteed the MBSs bought by the Fed, so they are nearly perfect substitutes with T-bonds.
Mike, No, this is completely standard use of “ceteris paribus.” DDs are endogenous in the model.
As far as the claim that banks will expand deposits at the same rate, that’s where the ceteris paribus assumption comes in. In the real world of course deposits may expand at a different rate, but that would be due to factors unrelated to the increase in the base.
6. October 2013 at 16:17
Scott
“I don’t understand what you mean by “base neutral”. Reserves are part of the base.”
That was a typo on my part; in fact, I meant QE as neutral with regard to economic activity (ie, sitting in ER). Earlier, you said of QE: “No. It’s sitting in the banks as excess reserves.”
Given that QE3 is $85B / month… excess reserves grew at a monthly average rate of ~$56B from Oct 2012 (QE3 Implemented) thru May 2013 (final month of ER data available from FRED). What are your thoughts on this analysis relative to your statement about QE sitting in excess reserves? If not the specific numbers, then conceptually the comparison of ER growth to QE volume?
Additionally, I am inclined to think that each additional dollar of QE is less ‘stimulating’, if you will. Evaluating growth of ER over the same time-frame as above, ER did grow at an accelerating rate (decreasing marginal utility) over time; this is not surprising to me.
6. October 2013 at 16:29
Scott,
This is my first activity on your blog, would like to say that I think it is fantastic and thought provoking. I really admire how so accessible you are that someone as irrelevant to these monetary policy affairs as myself can actually engage you on the subject; it serves as an inspiration to become more involved in something of great interest to me.
Keep up the great work!