There seem to be some misconceptions about the monetary consequences of actions that the Federal Reserve has taken to address liquidity needs.
One hears a fair bit of chatter these days along the lines of
The Federal Reserve (and other banking institutions around the world) have poured a couple hundred billion dollars into the system in an attempt to ensure that there is enough “liquidity” for banks to make loans
and references to
the Fed printing money and increasing the money supply on a high scale as if it was dropping money from an helicopter, thus the nickname of Fed Chairman Ben “Helicopter” Bernanke.
I hope in this post to get beyond these sound bites, beginning if I may with some details of the process whereby money is created in the United States. Where did the cash in your wallet come from? Presumably you got it from your bank or ATM. And the reason that the bank was willing to give you that cash was that you already had deposits in an account with the bank, which were in effect credits to obtain cash when you wanted it.
And where did your bank get that cash? If it is a member of the Federal Reserve, your bank got it from a Federal Reserve Bank. And the reason that the Federal Reserve was willing to give your bank the cash was that your bank had deposits in an account with the Fed, that give it credits to obtain cash when it wants it. These credits are known as Federal Reserve deposits, which I’ll simply refer to here as “reserves” for short.
And where did those reserves come from? The standard mechanism is that the Fed purchased Treasury bills from independent securities dealers, paying for them by creating new reserves in the dealers’ banks. The Fed now is the owner of the Treasury bills, and those banks now have new reserves, which they could use to obtain new dollar bills, if they desired.
To follow what’s happened over the last 6 weeks, it’s necessary to add a few details to this basic story. For day-to-day fine-tuning of interest rates and the money supply, the Fed usually does not use outright purchases of Treasury securities, but is more likely instead to rely on repurchase agreements, or repos. A repo is a short-term loan, often overnight or for just a few days, made by the Fed through a private securities dealer, which the Fed again provides by creating new reserves at the dealer’s bank. As collateral for the loan, the Fed temporarily takes possession of high-quality securities. A few days later, the securities are returned to the original owner and the reserves come back to the Fed. Although its effects are strictly temporary, on any given day the Fed’s outstanding repos have created reserves and thus potential dollars in circulation.
| Week ended Aug 8 | ||
| Total dollars created | 818,498 | |
| Treasury securities | 790,814 | |
| Repurchase agreements | 18,571 | |
| Less: Reverse repos | -31,647 | |
| Discount window loans | 251 | |
| Other | 40,509 | |
| Total dollars held | 818,498 | |
| Reserve balances | 5,447 | |
| Currency in circulation | 813,051 | |
The table at the right summarizes the Fed’s balance sheet (from Fed release H.4.1) for the week ended August 8, just before the summer fireworks in international capital markets. At that time, the Fed held $791 billion in Treasury securities and $19 billion in repos. The Fed can also take the reverse side of a repo, temporarily borrowing from private lenders and thereby absorbing reserves from the banking system. To calculate the total number of potential dollars in circulation at any point, we’d have to subtract this number from the Fed’s Treasury and repo holdings.
In addition, a bank that wants reserves can borrow directly at the Fed’s discount window, where the bank temporarily surrenders high-quality assets in order to obtain more reserves. On August 8, such borrowings were quite tiny.
There are some more details of the Fed’s balance sheet and classification of reserves that played little role in recent developments, so I’ve lumped these together in an “other factors” category in the table above. When you add together the Fed’s Treasury holdings, repos, discount loans, and other factors, and subtract off the reverse repos, you get the total number of potential dollars created, which came to $818 billion on August 8. Of that sum, only a tiny fraction ($5 billion) was held as reserve balances, and all of the rest had ended up as money in circulation. If you took the view that another $1 billion in Treasury securities purchased by the Fed will eventually mean another $1 billion in currency in circulation, that would be an excellent summary of the long-run reality.
Beginning August 9, the Fed aggressively used repos to add new reserves. Much of this was done with overnight repos, meaning that in order to keep the reserves in the next day, the Fed would need to conduct a new repo. The hundred billion dollar figure that some people use comes from adding together each day’s repo operations, which is a completely nonsensical calculation. At the height of these operations (the week of August 9-15), Fed repos created an average of $18 billion in new daily reserves compared with the previous week. It is also inaccurate to refer to such repo operations as any kind of bailout, since the Fed was physically holding securities whose value equaled that of the reserves temporarily injected.
 |
And what exactly happened to that $18 billion while it was in the banking system? The answer is– absolutely nothing. Banks simply held these funds as excess reserves, with nobody withdrawing a single dollar bill. This tremendous increase in banks’ desires to hold reserves beyond the amount that they were required was one of the remarkable aspects of this situation and the primary reason that the Fed needed to conduct such repo operations in the first place.
| Week ended Aug 8 | Week ended Aug 15 | Week ended Sep 19 | Change from Aug 8 to Aug 15 | Change from Aug 8 to Sep 19 | ||
| Total dollars created | 818,498 | 836,308 | 815,476 | 17,810 | -3,022 | |
| Treasury securities | 790,814 | 790,655 | 779,636 | -159 | -11,178 | |
| Repurchase agreements | 18,571 | 36,286 | 30,179 | 17,715 | 11,608 | |
| Less: Reverse repos | -31,647 | -31,357 | -35,735 | 290 | -4,088 | |
| Discount window loans | 251 | 271 | 2,421 | 20 | 2,170 | |
| Other | 40,509 | 40,453 | 38,975 | -56 | -1,534 | |
| Total dollars held | 818,498 | 836,308 | 815,476 | 17,810 | -3,022 | |
|---|---|---|---|---|---|---|
| Reserve balances | 5,447 | 23,905 | 5,561 | 18,458 | 114 | |
| Currency in circulation | 813,051 | 812,403 | 809,915 | -648 | -3,136 | |
And what’s been happening since August 15? As the table above indicates, the Fed continues to do a lot of repos, giving those pundits who want to add each day’s operations together some marvelous big numbers to play with. Although no longer at the same level of mid-August, the Fed’s repo holdings are still about $12 billion higher than they were the week of August 8. However, since then the Fed has also allowed some of its Treasuries to mature without rolling over. When the Treasury paid the Fed for these maturing Tbills, that ended up absorbing reserves in an amount that almost entirely offsets the added repos.
And how about discount window borrowing, which as of last week was still supplying a few billion in reserves in the system? This was more than offset by a $4 billion increase in reverse repos and some other factors, which drained considerably more in reserves than discount borrowing has added. The combined result of all these effects is that reserve balances are now right back where they were on August 8, and currency in circulation is in fact $3 billion lower than it was. The Fed has done its job, and the kvetchers have done theirs.
This is not to insist that concerns about higher inflation are unfounded. But, if one wanted to motivate such concerns from a monetarist perspective, one could not point to money that has been printed so far. Instead, the story would have to be that, in order to achieve the path for the fed funds rate that the Fed is now likely to set for the following year, the Fed will eventually need to add more reserves that do end up as more cash in circulation. In this scenario, markets have been reacting to an anticipation of future money creation and not to something that has already happened.
Now, I realize that this explanation may not be as entertaining as the whole Helicopter Ben meme. But on the other hand, it may be a whole lot more accurate.
Technorati Tags: macroeconomics,
Bernanke,
Federal Reserve,
inflation,
money supply,
economics,
money creation
This is one of the best, most illuminating blog posts I’ve read in several weeks. Superb, as usual, professor.
I really like that you generate valuable content in your blog. It takes so much more work than parroting, but it makes the blogosphere a place worth visiting.
Thanks for your writing on Econbrowser.
Number crunching is much more work and far less fun than writing fiction, and I’m sure I speak for most readers when I say that we appreciate you taking time out of your day to provide us with the facts.
Money creation and the Federal Reserve
…
Isn’t there another piece to money creation, the loans the banks decide to make? And, how many times their reserves can to commit to lending? I’d like to hear more about this.
What you’re completely overlooking is the amount of money created by the banking system. Broader money supply measures are growing very rapidly, which is reflected in the overvalued stock market, the soaring commodity prices and falling dollar. Since you made a reference to a post by Mike Shedlock, I would like to refer to my reply to that post:
http://stefanmikarlsson.blogspot.com/2007/09/how-should-money-supply-be-defined.html
Regarding inflation, the real issue is not that the Fed took action to stabilize the financial system in response to the subprime mortgage crisis. Rather the issue is whether or not lowering the target rate 50 basis points is appropriate. I think it was, but reasonable people can disagree about this.
That’s an excellent point, Stefan, but I was worried my post was getting a little long as is. Hopefully I’ll have an opportunity to discuss inside money another time.
Professor-
There’s a lot to digest in your post, but I thought your explanation was excellent.
It is not what the central bank was doing prior to 18 September 2007 that alarmed the bond, currency and commodities markets.
It is what the central bank did ON 18 September 2007 that set off the alarm.
The new set of incentivizations embraced by the central bank is tantamount to a message to the “American people” that they are saving too much.
Say what!
As we all know, the problem with the American people (among others) is that they save far too little or actually dissave.
The markets have heard the message loud and clear … growth at all costs is the goal, inflation is the tool.
BB wants to set off on another wild ride and the deployers of capital are taking protective decisions.
Nobody wants to be fooled again. Again.
Nobody.
Conceding the general accuracy of your description, there are two flaws: (1) the Fed took as collateral bonds of questionable value, and (2) the Fed was willing to roll over the repos as long as they deemed necessary. Together, that amounts to making a loan based on doubtful collateral. If the bank became insolvent during this period– as Nouriel Roubini is warning is the central problem in this crisis– the taxpayer is on the hook for the bad bonds. Plus, as Chrerick notes, we were offering fire sale prices for money, given that it was so thinly collateralized.
Stefan’s point is a good one. M3 is no longer reported. This is the measure that includes repos. The lack of transparency here feeds fear. However, the fall in dollar price is not primarily related to recent Fed actions, and the market is “soaring” only if one neglects inflation and currency effects. The dollar drop seems to me to be more closely related to fiscal policy, with a steep drop off somewhere around the time of the tax cuts and the declaration that the only wartime sacrifice Americans would have to make would be to shop more.
How does the Federal Reserve Bank value the “high-value” securities accepted as collateral. Hypothetically, wouldn’t it be unsustainable, if the primary securities dealer, such as Countrywide Financial, uses the increase in reserves to cover a run on assets without using the additional funds to make a productive investment? If the Federal Reserve Bank rolls over the repo to that institution, wouldn’t that constitute a feedback loop that ultimately will permanently expand monetary supply and leave the Federal Reserve Bank holding the “high-value” securities (i.e. — AAA rated CDOs)?
Please, forgive my 3rd-level conditional grammar mistake and my punctuation error.
Some of what follows may be presumptuous. I don’t mean to offend.
I don’t know what the pundits were saying, but my impression is that a lot of the drama earlier this month was related to the reality that some financial institutions required special attention because they found it difficult to come up with the daily balance that they are required to maintain with the Federal Reserve.
I think it’s extremely important that your readers come away with some understanding of the important, delicate and necessarily discrete things that the Federal Reserve can do when individual banks find themselves in difficult situations.
I think that you might want to continue your very good explanation with some observations on why there was legitimate interest in Federal Reserve discount window activity, and why the Federal Reserve was apparently in some cases allowing financial institutions to secure their reserve deposits with riskier securities than those you mention.
Your very good explanation of how the Fed makes cash available to the banking system seems like it would lend itself very well to some explanation of the “universe” which many financial institutions inhabit, a universe which includes large numbers of daily requirements to disburse “cash” to their customers and other parties and a beehive of hopefully profitable activities that enable the institution to successfully meet these obligations on a daily basis.
This becomes important if you believe that at least some of the press was appropriately relaying the concerns of responsible people that a “liquidity crisis” was possible; you can develop your discussion of the activities of individual financial institutions in a way that very clearly reveals the intricate inter-relationships of the various institutions and the ways in which “bad weather” in an individual institution can cause serious problems in other institutions in much the same way that very bad weather in one airport can cause real problems in many airports.
I’m not qualified to write this sort of thing, there’s a good chance that I’d get a lot of things wrong. But, continuing to press forward without shame, I’d really like to see you develop your explanation along the lines you have started in a way that includes:
1) Some discussion of how banking changed in the last 35 or so years. I would explain that while banks once had an enormous business lending short-term money to their non-bank customers like large manufacturing companies that needed cash for payrolls and to pay their suppliers, the availabiity of inexpensive computing and simple tools like spreadsheets and mathematical tools that enabled companies to price certain kinds of risk made it possible for those (non-bank) companies that had extra cash to manage the direct lending of short-term cash to other (non-bank) companies at rates that were more competitive than the rates banks could charge. I think this is what most people call “commercial paper,” although I have an impression that really describes a part of these non-bank transactions.
2. Using the above as a starting point, elaborate somewhat on the legal/regulatory and tax/accounting developments which, together with inexpensive computing and improved financial algorithms made it possible to “securitize” assets, and how this securitization made it possible for banks and other financial institutions to develop assets that could be tailored to meet various circumstances in ways that can help both the banks and their customers more efficiently and less expensively try to make sure that they can meet their obligations to have cash on hand when needed.
3. Some explanation of what leverage is, how leverage can create profitable investments, with some examples of how leverage has become important in various kinds of transactions. I think it may be a good idea here to go into the ways in which the demise of Glass-Steagal made it possible for many banks to get involved in “merchant banking opportunities”, discuss the role of leverage in some kinds of merchant banking transactions and look at how overall bank and financial institution profitability affects an institution’s ability to raise “cash”, particularly in circumstances where investments have gone sour or appear to be going sour.
4. Some little treatment of the scale of various kinds of financial activities – for example: Equipment financing, residential and commercial mortgage financing, trade financing, financing business mergers and acquisitions. I can imagine your using this discussion to help your readers get some idea of the scale of residential mortgage financing. Your approach is particularly suited here, because homeowners with mortgages pay cash to satisfy their obligations, and your readers now have some insight into the ways in which it became possible for mortgages to become aggregated together and securitized. Your readers can understand how mortgages relate to leverage, and mortgages seem to me to be an excellent way of setting out some thoughts about excessive leverage and how prudent bankers may behave.
5. By keeping your focus on the need to satisfy certain obligations in “cash,” and by providing some of the recent history of banking that shows how banking was forced to evolve as earlier lending markets disappeared and new opportunities became possible with among other things the tremendous growth of securitized (and hence tradeable) assets, you can now provide some context for the reasons why fee-for-transaction activities have become a very familiar part of how banks stay profitable. This also can lead into some treatment the ways in which banks separate mortgage initiation and mortgage servicing, and this may also be a good place to observe that there are probably very many financial institutions that are holding and have been trading in various types of debt obligations that are in whole or in part secured by mortgages that have a good likelihood of default.
I think that if you’ve been able to somehow subtly keep people aware that it’s difficult to be a financial institution, and that these institutions are very, very carefully trying to balance risks to maintain profitability your readers will understand how easy it is for things to get very weird if you have assets that you one day realize can’t be priced because some of the underlying assets aren’t performing. They may not be so “toxic” that they are fatal, but they become particularly “toxic” if they are large assets that simply can’t be priced because they’ve been aggregated in ways that make it difficult to find out how risky the individual underlying mortgages are, and even more dangerous when there’s leverage involved.
6. You can then perhaps try to give your readers some picture of the ways in which large financial institutions may have had difficulties because they had invested in or had other kinds of financial relationships with other financial institutions that perhaps had sufficient exposure to investments that were sour that it became important for the Federal Reserve to become involved.
_________________
I can imagine that this treatment could extend over several weeks, and it occurs to me that you could post it on this blog, but you might also consider putting up a wiki with you and some other professionals and maybe some students acting as “moderators” – you’re certainly right that there’s a lot of misunderstanding, and it was really great to see you go out and try and cut through some of the nonsense that’s out there because we do seem to have a remarkably large number of people in the press who could use better editors.
If you go the wiki route, you might be able to enlarge the topic a bit. I’d be very interested in following what people think about the ways in which liquidity “cash” concerns are managed when the Federal Reserve is operating in an environment where it must balance liquidity “cash” in a setting where the players are other central banks and financial institutions that are typically dealing globally and are trying to manage their accounts in ways that will promote economic growth and stability in their countries. Here I’m thinking really of the nuts and bolts – what kinds of transactions occur – who buys and sells. When you observe, “A repo is a short-term loan, often overnight or for just a few days, made by the Fed through a private securities dealer” I hunger for more information – what are the rules, what is the history of this market, who have the players been, how profitable is it, how accountable is it – the usual stuff. And I have zip idea of the nuts and bolts of lots of other things that it would be interesting to see developed by people who are knowledgeable and patient and insightful explainers.
So thanks for the posting and the little comment box. Again, sorry if this is somewhat pretentious and presumptuous (and so inordinately long).
Whoops in the my long post.
In numbered item 6, I wrote:
6. You can then perhaps try to give your readers some picture of the ways in which large financial institutions may have had difficulties because they had invested in or had other kinds of financial relationships with other financial institutions that perhaps had sufficient exposure to investments that were sour that it became important for the Federal Reserve to become involved.
I meant to write:
6. You can then perhaps try to give your readers some picture of the ways in which large financial institutions may have had difficulties because they had invested in or had other kinds of financial relationships with other financial institutions that perhaps had excessive exposure to investments that were sour that it became important for the Federal Reserve to become involved.
I was pointed to this post from Brad DeLong’s blog.
After re-reading your post, re-reading the comments on your post, and reading some comments on Brad DeLong’s blog, I realized that my long post here may seem to be off-topic.
Here’s what I posted on Brad Delong’s site that may explain why there’s some confusion:
I couldn’t quite “get” the post.
The post is titled “Money Creation and The Federal Reserve”. It’s about misunderstandings of the “monetary consequences of actions that the Federal Reserve has taken to address liquidity needs.”
But the primer isn’t so much a primer that would lead one to understand a lot about monetary consequences, rather it seems like a primer that’s largely explaining to people who may misunderstand what liquidity is how the Fed goes about moderating liquidity.
But it never explains to this primer-audience that the Fed was doing this because there were some financial institutions that (apparently) couldn’t meet their reserve requirements.
And so, for this primer-audience member, it was difficult to see where he was going with statements like:
“It is also inaccurate to refer to such repo operations as any kind of bailout, since the Fed was physically holding securities whose value equaled that of the reserves temporarily injected.”
or, re the discount window:
“On August 8, such borrowings were quite tiny,”
again, where you have a primer that appears to be explaining the mechanics of Fed interventions and not a primer that’s explaining the monetary consequences of fed actions. I think he may be leaving out too many steps for the primer audience.
I wish he had made it very explicit that there was something significant happening, and, for example, pointed out that while the discount window activity may have been a small number, it may have helped a small number of institutions from very unpleasant stuff.
And my impression looking at the comments that followed is that some commentors were economists/grad students or industry folk and some were primer-audience people who may have come away with a somewhat mistaken understanding of why there was legitimate press interest, and why the quote he uses early in his post isn’t, at least on its own so obviously “chatter”:
“The Federal Reserve (and other banking institutions around the world) have poured a couple hundred billion dollars into the system in an attempt to ensure that there is enough “liquidity” for banks to make loans”
______________
So I posted some suggestions for how he might improve his primer for general audience types like me, and will cross post this commment on his blog.
La Deluge, my aim here was a rather narrow one, namely, to review the popular claim that the Fed’s actions of the last six weeks could be characterized as “printing money” or excessive currency creation. I find such a claim to be factually in error.
You are quite correct that there are a number of other important issues relevant to recent developments that also need to be discussed. I’m inferring that you are not a regular reader of Econbrowser, and may have missed our extensive treatment of some of these concerns. You will find discussion of many of the points you raise under the Topics — Federal Reserve sidebar on our main page.
Professor, you present a very nice rebuttal to the uninformed outcries that we read here and there about the Fed’s recent actions. Technically, you are correct about the effect of the Fed’s actions these past six weeks.
But, when I see this chart — MZM growing at 10% year over year — I suspect that something bad is afoot with our money supply.
http://research.stlouisfed.org/fred2/series/MZM/chart?cid=30&fgid=&fgcid=&ct=&pt=&cs=Medium&crb=on&cf=pc1&range=Max&cosd=1980-11-03&coed=2007-09-10&asids=+%3CEnter+Series+ID%3E
The Fed is not maintaining a sound dollar. As Ron Paul pointed out to Dr. B. this week, such is immoral.
And, Dr. B., bring back M3; spend the $250K per year that it costs to produce the series.
Le Deluge,
With this:
“But it never explains to this primer-audience that the Fed was doing this because there were some financial institutions that (apparently) couldn’t meet their reserve requirements.”
It may be so because that idea gets readily Headlined into: “Banks Insolvent!”
Certainly, though, I agree with jg’s sentiment: “Professor, you present a very nice rebuttal to the uninformed outcries that we read here and there about the Fed’s recent actions. Technically, you are correct about the effect of the Fed’s actions these past six weeks.”
JDH,
John Hussman has a nice article making essentially the same points as you.
Yet, I think that the point that Stefan raised above about the money supply is important.May be the growth in money supply is what people refer to when they talk about Helicopter Ben? I am not sure if Bernanke has any control over it as long as government vastly overspends at the same time when foreign governments are willing to purchase US debt at a ver low yield. I would really appreciate it if you could shed some light on this.
Prof Hamilton,
many thanks, that was very illuminating indeed.
JDH:
Since you understand that the Fed never issues a dollar without getting a dollar’s worth of securities in exchange, I have to ask if you believe those assets matter to the value of the dollar? If you do, why would you give credence to fears of higher inflation? If every new dollar is backed by a dollar’s worth of securities, then the Fed’s assets automatically move in step with its issue of money, and there is no inflationary pressure.
Mike Sproul, certainly I regard money creation, which replaces interest-bearing government liabilities (Tbills) with non-interest-bearing government liabilities (money), as potentially inflationary. My point is there has been no money creation associated with recent developments.
I am not quite sure I buy your explanation for money creation in total. If not at at all.
What if there were no Treasuries? Let’s assume the gov’t had run at a surplus for quite a while and whatever treasuries that were in existence, were already bought by the Fed. How does it create money now? I am a afraid somebody who read your post would be painfully lost at this point.
In fact, I believe it is the overnight window that generates money. Because here, money truly enters the system. Money that didn’t exist before is created. Of course, money is destroyed when the overnight interest is paid. However, the money taken out generally exceeds the amount put in, and therefore there is a net growth of money supply.
The Treasury mechanism is indeed a way to increase liquidity, but it does not create money.
I hate to disagree with a professional, but I see their is real confusion about this issue, and it is an important one.
A former FOMC member blogs on the Fed’s activities:
http://www.bob-mcteer-blog.com/the-feds-tool-kit/
Vorpal, in any regime you need a mechanism to get newly printed money into people’s hands. Barring helicopter distribution, you would do that by using the newly created money to pay for something. A really terrible mechanism is to use the newly printed money to pay the government’s bills. The reason that is terrible is that it gives the government an incentive to print more and more. Every stable economy instead uses essentially the mechanism I described above for the U.S., namely, using newly created money to buy assets of like value in arms-lengths transactions. In the U.S., the overwhelming majority of such assets purchased happen to be Treasury bills. This is simply a statement of fact. There should be nothing the least bit controversial about any of the claims I make in this post.
As for the suggestion that the discount window is the true source of money creation, this is beyond silly. As the tables above show, as of August 8, there was $250 million in outstanding discount window loans and $818 billion in outstanding cash.
If the Fed owns the TBills then does the US gov’t pay the Fed when they mature? If so, doesn’t that take money out of the system?
If the economy grows at around $1 trillion a year, then that implies that ~$20 billion is added a week. Given the multiplier effect, isn’t that a net borrowing of around $1 billion a week?
The only number I see near that figure is the repos.
As long as the TBills need to be paid, then I don’t see that adding net money to the system. There must be more to it.
Moreover, if the overnight window only affects a paltry $250 million in a $13 trillion economy, why do so many investors think it is such a big deal? That makes no sense.
I’m just not buying your explanation in full. Not that you don’t understand it, but you haven’t communicated what I consider to be a cohesive picture.
OK, Vorpal, first let’s be careful about using the term “money.” In this post, I am focusing on the green stuff you carry in your wallet, i.e., currency. There is a certain amount of the stuff that has been created at any given point in time. We know what that number is, because the government printed it and they keep track. That number is about $800 billion. It is what economists call a “stock” variable, in that it is measured at a point in time, like the total amount of water in a lake at a particular time.
When you talk about a 13 trillion dollar economy, you are talking about something else. You are talking about the total value of final goods and services produced over a particular period, which is a flow variable. You can’t calculate that number at a point in time, you can only calculate it over an interval of time, for example, so much produced each month, or so much produced each year. Think for example about how much water flows through a river during the course of a month or a year.
There’s no particular reason these two numbers have to be related. For one thing, they’re measured in different units. A stock is measured in dollars and a flow is measured in dollars per year. A given dollar bill over the course of a year might pass from your hands to somebody else’s and then who knows where. Furthermore, not all transactions are paid for with dollar bills. So it’s completely wrong to think the total supply of money and the total value of annual GDP ought to be the same or even comparable numbers.
Getting back to the stuff I am talking about, green currency, as I say it is really a very straightforward question to describe how that stuff got printed and ended up in your wallet, and that is exactly the process I’ve described for you here.
You ask all sorts of other questions, all of which I could perfectly well answer, but let me suggest that your main confusion comes from not distinguishing between physical dollar bills and other questions of interest. The important thing about physical dollar bills is they don’t get created or destroyed when I pay you or you pay me. If I pay you with cash, my cash holdings go down and yours go up but the total cash is the same. If I don’t pay you with cash, the total cash is the same. The total cash is controlled by the process I describe and not by any of the other things you are thinking about.
Vorpal,
Take a look at the previous posts JDH has provided on how the FED works. I think it might help you understand the process of money creation.
Great post professor. There is so much bad information out there it is no wonder that people are confused. Much of the bad information comes right out of the mouths of economic commentators and some economists themselves.
Professor,
Nice explanation. How often, and under what circumstances, does the Fed buy gold or other commodities instead of Treasury bills?
To my knowledge, TedK, the answer is never.
JDH,
One of your best posts. Quite helpful, and much appreciated.
In the near future, its possible that currency growth will ramp as the Fed finances a ballooning Federal deficit (mortgage bailouts, state budget bailouts, pension bailouts, unemployment insurance payments, etc). Perhaps this is what the dollar and commodities markets are really discounting. Bernanke’s famous “helicopter” speech specifically cites Fed financing of a growing deficit as a means of fighting deflation.
As a UCSC econ undergrad, I’m almost certainly throwing my weight around more than I should.
However, I believe theres some slight fault, not with your explanation, but the generalization of the repo securities the FED ultimately purchased.
You said:
“It is also inaccurate to refer to such repo operations as any kind of bailout, since the Fed was physically holding securities whose value equaled that of the reserves temporarily injected.”
and, more importantly:
“As collateral for the loan, the Fed temporarily takes possession of high-quality securities.”
These are both odd statements, as the large cash injection by the FED recently (august) was 38 billion (19, 16 and 3 billion), all collateralized with mortgage backed securities.
The FED couldn’t sell an MBS onto the open-market even if it could legally. No one wants them, which brings us to the unfortunate truth: those securities are not worth the 38 billion the FED paid for them. Right now, they’re incredibly risky, certainly no where near the quality of a T-bill. So much for quality!
Would you rather have 38 billion in cash or in mortgage backed securities? I, personally, would take the money.
I’m assuming the FED, in a repurchase agreement, isn’t stuck with the asset. Eventually it will get the money back when it stops renewing the repos.
In any case, I find your implication that this was somehow a fair deal because the Fed only accepts “high quality” to be inaccurate in this case. The Federal Reserve took on a terrible deal by accepting those securities as collateral. This was an extremely generous move, and I can only assume that it was done as to save the hide of whomever was holding onto one of these abysmal mortgages.
Professor, you are talking about the green stuff that the Fed has created or is/isn’t creating, which is about $800b. The bank reserves of this stock amounts to $49b, and that required amount hasn’t been raised in many years, although M3 has increased enormously. True?
Obviously M3 has grown at a ferocious rate, the last I heard at 14%/yr. It seems to me that banks and other “financial institutions” have themselves created huge amounts of money by means of leveraging. And it is this stock that now is threatened, not the stock of green stuff.
That is where the recent problems arise, because that new money has been created outside the realm controlled by either mechanisms of the Fed or by regulators. And there is little now that can be done to address the problem that the collateral underlying much of that new stock is not what it was assumed to be–i.e. solid. It is dropping in value, perhaps by 5%, perhaps by much more, and that percentage of M3 is disappearing from the US (and global) economy. Due to leveraging, we may be talking about tens of trillions of dollars.
My point is this: to prevent this loss of money from tailspinning the US economy into a terrible depression, the Fed must *replace* that funny money with real money. Liquidity and repos will not do the trick, by your analysis. One way or another, it seems to me, the Fed and Treasury will have to create new debt and monetize it–if they are not doing it already through means that you are overlooking.
If I completely misunderstand the situation, you can say so. If my mistake can be easily pointed out, I hope you will take a minute to do so.
Kevin, I agree that I would prefer the Fed not use top-tranche MBS for repos. But what I think you and several other commenters above may be overlooking is the fact that the very short-term nature of the repo inherently limits the risk. Let’s look specifically at the $38B in operations on Friday, August 10 that you mentioned (details of which can be found here). These were all 3-day repos, meaning the money had to be repaid on Monday, August 13. The total repo operations on Monday, August 13 only came to $2B. I do not know that any of that $2B would be part of the original $38B– in general, the Fed does not “roll over” the identical repo. But even if it had, $36B was repaid in full on that Monday.
Kevin Adolph:
In any case, I find your implication that this was somehow a fair deal because the Fed only accepts “high quality” to be inaccurate in this case. The Federal Reserve took on a terrible deal by accepting those securities as collateral. This was an extremely generous move, and I can only assume that it was done as to save the hide of whomever was holding onto one of these abysmal mortgages.
The MBS the Fed accepts in repos are agency-backed MBS (Fannie, Freddie). These are not just high quality, but extremely high quality. I haven’t heard of any problems in the market for these securities.
Unirealist, I last discussed M3 here; perhaps I should take up this issue again.
After reading much more from various authorities, including the Fed itself, I conclude that the repos are the real source of money. My understanding from multiple sources, the repos are similar to the overnight-window (very short term), except they are secured.
The purchase of securities is a good way to regulate liquidity, but unless the Fed accrues (net) more securities each year, then the money supply will not increase through security acquisition.
As for the comments on the flow of money into the system, relating to GDP, I don’t think they were that far off either, at least according to the graphs at this page: http://en.wikipedia.org/wiki/Money_supply
The graphs show that the the growth in money supply and the growth in GDP are on the order of 1:1 . This was the premise of my calculation.
In the end, my basic point of contention, is that the growth in the money supply (money creation) occurs via loans by the Fed, either through the discount window, repos, or some other lending process. This is fundamentally different than selling/buying securities.
Maybe I’m wrong again, but I haven’t read anything by anybody that is inconsistent with the above.
Vorpal, the number of net Treasury securities held by the Fed does indeed increase each year, essentially by the same amount that currency in circulation increases each year. The dollar value of the Fed’s repo holdings does not increase each year. A 1-day repo comes back to the Fed the next day. The currently outstanding stock of repos is less than 4% of current Fed holdings of Treasury securities. The Fed’s currently outstanding stock of repos less the Fed’s currently outstanding stock of reverse repos is a negative number.
Calculating the number of dollar bills that the government has printed and how they got into circulation is not a tricky, confusing, or controversial business. You imagine a difference of opinion among experts on this matter that simply isn’t there. The numbers are reported in Release H.4.1 and all I have done is explain how to read them.
Professor,
thanks for the enlightning article. Perhaps you could ‘tie’ together the M3, Discount window (this one), FFR, dollar value discussions into an Inflation article. Somewhere some how everything is more expensive and no one can explain why(wait unil the exporters to the US are unable to maintain their ‘pegged’ exports to the dollar as has been recently reported by the FR themselves and we can actually experience the inflation we all speculate about). I think having all these abstract ideas seperately serves to confuse the uneducated like myself. Perhaps a discussion that delves into the interelatedness of these seperate mechanisims and how ‘inlation’ occurs from someone with your understanding of them would be helpful.
I see your point. So their securities do grow each year. That’s that. The money is being lent..to the US govt…in ever increasing amounts.
I still have to think through the repercussions of that. On the face of it, it seems that the taxpayer is getting hurt, while TBill holders are getting the benefit, but I would have to think that through.
Thank you JDH, I think I have a better insight into the process.
JDH
“certainly I regard money creation, which replaces interest-bearing government liabilities (Tbills) with non-interest-bearing government liabilities (money), as potentially inflationary.”
On day 1, a bank receives 100 oz. of silver on deposit and issues 100 paper receipts (“dollars”) in exchange. On the next day, the bank issues another $200 for an IOU worth 200 oz.
It sounds like you believe the $200 is potentially inflationary. But if the dollar fell in value to .99 oz., the bank could sell the IOU for 200 oz, spend 198 oz to buy back 200 of its own dollars, and earn a free lunch of 2 oz. That can’t be, and there’s no reason why it should be. The $200 is adequately backed by the IOU, so there is no inflationary pressure. Certainly, if the bank issued $200 without getting any new assets, then the dollar would fall, but that’s not the way this bank operates, and it’s not how the Fed operates either.
I’m not sure I understand your analogy, Mike, but it is certainly true that a monopoly on the ability to print money gives the government an opportunity for a free lunch. What drives this economically is the fact that money has social value as something that facilitates exchange. That social value makes me willing to surrender a real potato to the government in exchange for a piece of paper– that paper is so much handier to buy a shovel with than is a potato. With a growing economy, we have more exchanges between shovels and potatoes that we’d like to make every year, and at a fixed dollar/potato exchange rate we’d be willing to give some more potatoes to the government each year to get those extra desired dollars. But, if the government prints money faster than our economy grows potatoes, over time the number of potatoes you could get with a dollar will fall.
Professor, I was surprised to see the large and increasing volume of reverse repos. Since the Fed has been quite active in using repos to provide additional liquidity to the banking system, the two items seem to be at cross-purposes. Since I was curious, I went to the H.4.1 release and found that the reverse repos are all being done with “Foreign official and international accounts”. Do you have any thoughts on who the reverse repo counterparties are and why the Fed would be increasing their reverse repo activity and soaking up liquidity at the same time they are increasing repo activity and adding liquidity?
Good question, Peter. Can any readers help us with this one?
If the reverse repos are all being done with “Foreign official and international accounts”, then I think Fed is smart. They are borrowing from foreign countries to add more liquidity to domestic market. Meanwhile, they are trying to keep the Dollars available in global market at certain level, so that dollar won’t devaluate too fast. Fed is passing more credit risk into foreign countries.
I think not, Sha Lou. This is a tiny number, and may have something to do with the separation of the international desk and the open market desk at the Federal Reserve Bank of New York.
Thank you for taking the time to clarify, professor. This thread has been extremely illuminating!
Professor, I really like your posts on the Fed. Best on the net.
I do have a few issues. You write that: “A repo is a short-term loan, often overnight or for just a few days, made by the Fed through a private securities dealer, which the Fed again provides by creating new reserves at the dealer’s bank. As collateral for the loan, the Fed temporarily takes possession of high-quality securities.”
From my reading on this subject, a repo acts as a loan, but legally is treated as a purchase/sale transaction. This quibble is actually important because certain sections of the Federal Reserve Act govern the Fed’s power to engage in purchase/sale transactions, while other sections govern collateralized lending. Specifically section 14 applies to purchases/sales, while section 13 and several others apply to lending. If, as you say, repos were actually loans, the Fed would be allowed to buy far more than just government securities via repos. They would be able to buy the same securities they are allowed to discount, including sub-prime MBS. But we know this is not the case.
A good legal source that talks about this distinction is:
http://staging.sifma.org/regulatory/briefs/OrangeCounty.pdf
You also wrote: “In addition, a bank that wants reserves can borrow directly at the Fed’s discount window, where the bank temporarily surrenders high-quality assets in order to obtain more reserves. On August 8, such borrowings were quite tiny.”
Your decision to focus on August 8 here is a bit misguiding. By Sep 12, outstanding discount loans hit $7 billion. And though it is easy to write this off as only $7 billion, that amount is the highest since 9/11 and before that the S&L bailout.
See this chart:
http://research.stlouisfed.org/fred2/series/TOTBORR?cid=122
I was wondering if you could answer a question that has been stumping me. Say the Fed loaned out money via the discount window to a certain bank, accepting suspect collateral as security. If that bank went bankrupt before the loan came due, and if the value of the collateral had fallen to zero, what would the Fed do? It would have issued money into the economy, but that money would be unbacked since the collateral it had accepted was now worthless. Any thoughts? Could you explain this using the Fed’s balance sheet?
One other thing, on the asset side of your Fed balance sheet you left out gold ($11 billion) and SDRs ($2 billion). This would lift the value of the supply and disposition of potential dollars to around $833 billion. This puts the Fed far in excess of currency circulating of $813 billion plus reserve balances of $5 billion, no?
JP, gold, SDRs, and a number of other items are grouped into the “other” category that I created. Currency is exactly the number reported.
The Fed does accept securities other than Treasury securities for a repo.
Beginning the analysis with August 8 is in my opinion not the least “misguiding”. I begin the analysis there because that is just before the aggressive actions of the Fed that some have criticized were implemented. I describe what happened that week and the cumulative consequences of everything that has happened between then and now.
As for your hypothetical question about worthless collateral, certainly the cash created would stay in the system, so as a matter of accounting one would have to increase my “other” category so as to balance the loss in repo value. Technically if you refer back to the original H.4.1 balance sheet, the “Other liabilities and capital” entry would have to be decreased (and perhaps become a negative number) by the amount of the loss. Since this is subtracted from other assets to create my “other” category, this would show up as an increase there. Currency outstanding would be unaffected by the Fed’s capital loss, so in effect the consequences would be inflationary. The Fed could offset this by selling securities to absorb the reserves back in.
Profesor Hamilton, thanks for the reply. I understand now how you got the “other” category now.
“The Fed does accept securities other than Treasury securities for a repo.”
I agree with you, it does. It accepts agency debt and agency guaranteed MBS. But they are only allowed to do this because section 14 of the Federal Reserve Act states the Fed can…”buy and sell in the open market, under the direction and regulations of the Federal Open Market Committee, any obligation which is a direct obligation of, or fully guaranteed as to principal and interest by, any agency of the United States.”
I think my point still stands that repos are purchases/sales, and not collateralized lending. Like I said, I don’t want to quibble but the difference in language has very real implications.
As for the hypothetical situation, you said that the Fed must decrease its “Other liabilities and capital” entry. How would it go about doing so? I don’t want to misunderstand you here, but are you saying it could do so by selling securities?
JP, there seem to be two different questions you are asking. One question is about accounting: How would a capital loss be entered on the Fed’s balance sheet? The answer to the accounting question is, you would reduce repos on the asset side and and reduce net equity on the liabilities side each by the amount of the capital loss, with all other entries the same.
Perhaps instead you’re asking, What would be the immediate consequences for outstanding currency? The answer to that question is, nothing. The currency would still be there, and wouldn’t flow back to the Fed when the loan is repaid because, under your hypothetical scenario, the loan is not repaid and there are no funds to cover it.
If the Fed didn’t like that outcome in terms of currency, namely, they wanted the currency to flow back to the Fed, they would then conduct a new and separate operation– sell a Treasury bill. This would correspond to a new pair of entries to the Fed’s balance sheet, which are added to the original changes to repos and net equity described in my first paragraph. Namely, the new entries would be, Fed holdings of Treasury securities go down and currency goes down.
Every transaction enters twice on a balance sheet; it’s balanced before the transaction, and it’s balanced after.
JDH
“I’m not sure I understand your analogy, Mike, but it is certainly true that a monopoly on the ability to print money gives the government an opportunity for a free lunch. What drives this economically is the fact that money has social value as something that facilitates exchange.”
If it’s not clear, 10 minutes spent reading about the real bills doctrine would probably make it clear.
Anyway, that free lunch I spoke of is available to anyone, not just a monopolistic central bank. If the dollar had fallen to .99 oz/$ because, as you said, there was more money chasing the same amount of goods, then anyone could sell 198 oz on the open market for $200. Then they could return that $200 to the bank and exchange it for the bank’s IOU, which can then be sold for 200 oz. The existence of this 2 oz. free lunch implies that the quantity theory of money is wrong. There is no such free lunch under the real bills view.
Professor Hamilton
“The answer to the accounting question is, you would reduce repos on the asset side and reduce net equity on the liabilities side each by the amount of the capital loss, with all other entries the same.”
This sounds to me like how a corporation takes a writedown. It takes a loss on its income statement, which has the affect of reducing the value of its assets on one side of the balance sheet and its shareholder’s equity on the other side.
So theoretically, the bank could reduce its “Other liabilities and capital” to below zero if necessary, as you said, entirely offsetting the value of the now worthless collateral that had been accepted for new money. I think I understand that point.
But what about the Federal Reserve Act Section 16 requirements that each circulating federal note be backed by any of the following: gold certificates credits, SDRs, U.S. government securities, federal agency securities, discount window borrowings, and foreign government or agency securities?
http://www.eagletraders.com/advice/securities/federal_reserve_notes.htm
Source: Encyclopedia of Banking & Finance (9h Edition) by Charles J Woelfel
Given this law, it’s not just a matter of restoring assets equals liabilities in the event of collateral becoming worthless. It is necessary that a certain type of asset (the above securities, held via either outright purchases, repos, and/or loans) be returned to a level equal to or greater than a certain type of liability (reserve notes in circulation).
If collateral becomes worthless as in our hypothetical, reducing repos on the asset side and net equity on the liability side via a writedown does indeed make Fed assets equal to Fed liabilities, like you said. But as far as I can see it does not satisfy section 16. It does not return currency in circulation to the same level as securities held via either outright purchases, repos, and/or loans.
Am I off the beaten track here? If not, how would it satisfy section 16?
Nobody would want to promise to convert money at a fixed rate of exchange to any real commodity over time, Mike. And by a government monopoly, I mean that no one can issue a paper obligation that functions as nicely for purposes of exchange as does the government-issued currency. We all know that next year a potato will cost $1.02. We nontheless happily surrender 1 potato to obtain 1 dollar today, just as we will happily surrender 1 potato to obtain 1.02 dollars next year. The reason is that it’s handier to have the dollar, even though it’s not as good at keeping its value as a potato.
JP, let me repeat that this all is exceedingly hypothetical. In my opinion, the letter and the spirit of the law are perfectly well satisfied by the procedure I described. One could also mention Fed interest receipts which could rebuild the negative net equity if that’s your big concern. I think you are comingling questions of accounting, economic policy, and legal definitions in a way that does not strike me as terribly productive. It would not be that big a problem for the accountant, the policy-maker, or the lawyer if there was a default on discount window collateral, though they might each be thinking about what needs to be done in slightly different terms. Furthermore, it is not very likely to happen.
Professor Hamilton
You’ve got me there, I am indeed trying to understand the Fed from as many viewpoints as possible. I wouldn’t describe this is as unproductive though, just an honest attempt to shoot for the big picture.
And yes, my hypothetical is extremely er.. hypothetical. But I think it makes sense to understand the Fed both in its day-to-day normality, and from the standpoint of what-if scenarios, no matter how unlikely they are to happen. I believe in asset management they call it stress-testing, trying to figure out what damage unlikely events will cause to portfolios.
Once again, thanks very much for replying to all my questions. I look forward to more posts on the Fed as yours are far better than the other economics blogs out there and I have learnt a lot from them.
Prof hamilton:
i read a nice chain going in there..but you stopped at FED buying the treasury securities. But from whom? Who creates these securities?
In the end i guess, we do come back to printing (electronic) money (but as long as it equals nominal GDP growth, everything would be ok)
JDH
“Nobody would want to promise to convert money at a fixed rate of exchange to any real commodity over time,”
Almost all central banks initially make their money physically convertible–that is, the bank agrees to buy back its dollars for a physical amount of gold, etc. Eventually, insolvency and the resulting bank run forces them to suspend physical convertibility, but virtually all central banks continue to maintain financial convertibility–that is, the bank agrees to buy back its dollars with a dollar’s worth of its assets. If a dollar is pegged at some arbitrary value (one ounce of silver), then that peg can be maintained either with physical convertibility or with financial convertibility. The Fed, for instance, maintains the value of the dollar using only financial convertibility, that is, by conducting ordinary open-market operations. The Fed could not do this if it did not have enough assets to buy back all of its dollars. Hence my claim that the dollar has value because it is backed.
As to a monopolistic central bank, that claim can’t be used when countries are small, weak, and close together, but standard monetary theory still claims that even the smallest countries are able to issue fiat money. If that’s true, and if money really can have value in excess of its backing, why has there never been an example of a central bank that did not hold assets against the money it issues?
Money Mambo
While I’m still a little fuzzy on the actual specifics of this explanation of just what, exactly, the fed’s been doing to avert various monetary crises, the conclusion seems to me sound enough. Those who get their economics reports…