Thursday, December 9, 2010

What Does “Printing Money” Mean?

When I heard Ben Bernanke on 60 Minutes Sunday, I was initially taken aback when he said that QE2 did not constitute “printing money.” Obviously it’s not physically printing money, but nobody ever uses that phrase literally. If creating bank reserves is not “printing money,” then what is? My first thought, maybe he means that the increase in base money is not expected to lead to an increase in bank deposits. But if that’s the case, why would they be doing it? It’s hard to imagine QE2 being effective without causing bank deposits to increase. Sure, they may not increase via the textbook money multiplier process, given that reserve requirements are not currently binding and reserve ratios are not expected to be stable. But that’s a cop out: the Fed is increasing bank reserves; it hopes this action will lead, by whatever process, to an increase in bank deposits. How is that not printing money?

Apparently it gets worse. I had forgotten what Bernanke said on his earlier 60 Minutes appearance, but Jon Stewart has a clip where, in reference to QE1, Bernanke essentially acknowledges that the Fed is printing money. Naturally, Jon Stewart was amused by this seeming inconsistency, as was I.

But I’ve been thinking about this a bit more, and I no longer think Bernanke’s statements are inconsistent. The problem is that the definition of the word “money” is not as clear cut as it seems. Indeed, one might argue that the whole concept of “money” is no longer useful (at least to economists) in a world where bank reserves pay interest and people pay bills with credit cards and with checks drawn on bond mutual funds.

Long, long ago, before I started graduate school, I used to think that the concept of money was fairly straightforward. There was cash (Federal Reserve notes, as well as coins), and there was money in the bank, and those were money. And OK, there were money market mutual funds, and those were “sort of” money. When I really started to think about it, I realized this framework was inadequate, especially given the concept of liquidity preference that I was trying to use, where the cost of holding liquid money was supposed to be the interest rate. (Money market funds pay interest, and they’re almost just like money, and even some bank accounts pay interest, so apparently you can hold liquid money without giving up the interest, so where’s the cost?)

At the time I had a solution: stop thinking of bank accounts and such as money and instead just think of “outside money,” money created by the central bank. This approach sort of seemed to work. If you wanted to hold “money” in this sense, you had to give up the interest, and a bank’s willingness to pay interest on deposits (no longer considered money) was influenced by the amount of actual money that the bank needed to hold in order to maintain that deposit. And policymakers actually had control over this kind of money, so the concept fit well with the simple assumption that the quantity of money is determined by policy.

But in 2008, the Fed started paying interest on bank reserves. (Some other central banks had already been doing so for a while, but, being a provincial American, I hadn’t really noticed.) To be honest, it didn’t occur to me at the time, but this change totally destroyed my concept of money. If bank reserves pay interest, then they aren’t money, because you don’t have to give up the interest in order to hold them. But surely they are money, because monetary policy consists primarily of manipulating the quantity of bank reserves. Epistemological fail!

So are bank reserves money or not? I don’t know. But here’s something to think about: what is the difference between bank reserves and Treasury securities? They both pay interest. They’re created by different institutions, but so what? They’re both ultimately obligations of the government: the interest paid on reserves comes out of the Fed’s profits which would otherwise go into the Treasury.

You might say that Treasury securities have to be paid back, while bank reserves don’t, but that’s merely a function of the banks’ willingness to hold reserves: if the banks (and their customers) want cash instead, the reserves have to be “paid back” to the banks/customers. And if Treasury security holders want to roll over their securities, then Treasury borrowing doesn’t have to be paid back. So there’s no real difference there. The maturity is different, it’s true: bank reserves have zero maturity, while Treasury securities have maturities ranging from one month to 30 years. But that’s not a fundamental difference: it just means that bank reserves are a special case, not that they’re a different kind of entity.

OK, bank reserves can be used to fulfill reserve requirements, and Treasury securities cannot, but again so what? Banks are subject to a number of regulatory requirements, which depend on various aspects of their asset structure. There’s really nothing special about the reserve requirement. Today capital requirements are closer to being a binding constraint than reserve requirements, so for most banks Treasury securities are just as good as reserves when it comes to fulfilling regulatory requirements. Maybe someday in the future banks will once again have a particular need for Fed-issued zero-maturity assets to fulfill a regulatory requirement, but I don’t see how that makes such assets a fundamentally different type. Show me junk bonds, T-bills, and bank reserves, and it seems to me that, if anything, the bonds are the fundamentally different type of asset.

I submit that bank reserves are essentially just zero-maturity government debt. You can call them “money” if you want, but the application of the term is pretty arbitrary. And if you’re going to call bank reserves money, why not call T-bills money as well? And for that matter, T-notes and T-bonds: those are just long-maturity money.

So back to Ben Bernanke. What is QE2? It’s an exchange of bank reserves for longer-maturity Treasury securities: both forms of government debt; the only substantial difference is the maturity. QE2 is not, in any important sense, “printing money” but merely a restructuring of the government’s debt.

And what was QE1? It was (largely) an exchange of bank reserves for private sector assets, essentially an exchange of government debt for private debt. It’s very important that QE1 involved “printing” substantial quantities of government debt that would not otherwise have existed. That government debt happens to be “money,” though, rather than what we usually think of as government debt. So, rather than go through a whole explanation of how bank reserves are really government debt, the simple and substantively correct way of explaining QE1 is that it constituted “printing money” in exchange for private sector assets.

DISCLOSURE: Through my investment and management role in a Treasury directional pooled investment vehicle and through my role as Chief Economist at Atlantic Asset Management, which generally manages fixed income portfolios for its clients, I have direct or indirect interests in various fixed income instruments, which may be impacted by the issues discussed herein. The views expressed herein are entirely my own opinions and may not represent the views of Atlantic Asset Management. This article should not be construed as investment advice, and is not an offer to participate in any investment strategy or product.


Niklas Blanchard said...

I believe this is pretty close to where MMT'rs (or horizontalists, or whatever) like Winterspeak stand on the question of "what is money".

Certainly, he views the central bank as only having the power to alter the term structure of debt maturity -- something he believes has no real macroeconomic impact.

Richard H. Serlin said...

"(Money market funds pay interest, and they’re almost just like money, and even some bank accounts pay interest, so apparently you can hold liquid money without giving up the interest, so where’s the cost?"

One cost, and a substantial one, is that the interest rate is a lot lower than what you could have gotten with less liquid vehicles.

David Andolfatto said...

Andy: I like to think of the fiscal authority as determining the total outstanding debt, and the monetary authority as determining its composition (historically, between intererst-bearing and non-interest-bearing debt). An OMO that increases the ratio of money/bonds held by the public can now be thought of as "restructuring the government debt." Of course, "printing money" is involved in this transaction; but really, who cares how we label it? It is more important to identify the precise operation.

Dmitri said...

My understanding was (as you pointed out) that QE1 was focused on lowering short-term borrowing rates, while QE2 is attempting to bring down longer-term (I think up to 7-8 yr horizon). However, I'm still not 100% clear on the distinction between printing money vs effectively restructuring government's debt and not printing money in QE2? Replies would be appreciated.

David Pearson said...

The effect of "restructuring" public debt is to keep risk-free real rates below equilibrium. The Fed's 2003 "extended period" language did the same thing by providing an arbitrage opportunity in maturity transformation by shadow banks.

There is no direct inflationary consequence to the existence of large Excess Reserves. Instead, inflation potential arises from that below-equilibrium real rate. Markets extrapolate that real rate into the future, and actors (particularly the fiscal authority) enter into contracts with that expectations in mind. The difficulty arises from eventually allowing that rate back to equilibrium, as it "shocks" actors that took on leverage into de-levering.

Take a look at a chart of real rates in 2000-2010, and you will see an economy unable to stomach "normal" real rates. Today, the fiscal authority depends on low real rates in the absence of robust growth. Without it, long term deficit projections are likely to be unacceptably high. If inflation accelerates with 2-2.5% growth, the Fed will be unlikely to act.

Tom Hickey said...

Think of "money" as a token for something. Non-commodity money has no inherent value, as does a gold coin, for example. Modern money, especially the electronic money in which most transactions occurs, is simply a token.

What is money a token of? The obvious answer is that money is a token for what it can purchase. It's value changes relative to purchasing power. This is what "inflation," "deflation" and "price stability" mean.

There is money, which is liquid, like cash, checks, and live credit cards, and "moneyness," which is somewhat illiquid, like a time deposit or Treasury security that needs to be made liquid before it can be used. But the distinction is largely blurred nowadays, since one can switch funds immediately from savings to checking online, unless the bank puts a hold on the time account (unlikely). So money is that which can be used to buy stuff, including other financial assets. Money is a token for these real things it can purchase.

What about bank reserves? Bank reserves are not tokens in the same way. They are tokens of tokens, in that they are used exclusively for settlement in the overnight bank market. When someone uses a credit card, check etc., — anything but cash in which the transaction is settled immediately on the spot — the transaction has to clear through the interbank market. Bank reserves are used by banks to settle up their accounts. When I cash a check, the bank debits my deposit account and uses reserves as a token of the this in the settlement system to clear my check with another bank or maybe the IRS.

Reserves are a tool that makes non-cash money work. Reserves are not "money" in the commonly understood sense, although they constitute "base money," without which settlement doesn't take place under the current arrangements. Reserves are actually a holdover from the days of convertibility, and they are now not "reserves" at all since there is no backing. They are just numbers on the FRS spreadsheet that keep track of settlement among financial institutions, including Treasury.

Failure to understand the what the FRS is an does and how the accounting works, as well as the vertical-horizontal relationship of money, and the hierarchy of money leads to great confusion, even among economists. This is what Modern Monetary (MMT) Theory sets about illuminating.

Musgrave said...

1. It is nonsense to describe the quantitative easing of short term government bonds as printing money because these bonds are regarded in the world’s financial centres as virtually the same as cash.

2. As to QEing longer term government bonds or private sector bonds, there is still no money printing taking place in that there is no increase in private sector net assets (assuming a fair price is paid for the assets). But you could argue that SOME money is created in that the relevant assets are made more liquid.

3. GENUINE MONEY PRINTING to my mind consists of an unfunded budget deficit. I.e. the government / central bank machine simply writes checks to fund its purchases, with no corresponding increase in tax or borrowing. The accounts of those who receive the checks are then credited at the central bank (or to be more accurate, the account of the check recipient at their commercial bank is credited, and the latter’s account at the central bank is also credited). In this case every $100 worth of check results in a $100 increase in private sector net (and liquid) assets.

It never ceases to amaze me that 99% of the population (economists included) cannot see the distinction between QE (i.e. swapping bits of paper of approximately equal value), “genuine money printing”. The Japanese spend a decade swapping bits of paper. That’s why they lost a decade. MMTers have grasped the distinction.

Richard H. Serlin said...

"So back to Ben Bernanke. What is QE2? It’s an exchange of bank reserves for longer-maturity Treasury securities: both forms of government debt; the only substantial difference is the maturity. QE2 is not, in any important sense, “printing money” but merely a restructuring of the government’s debt."

First, the Feds purchases of bonds can cause the creation of new bonds ("money"). If the Fed decides to buy a trillion in corporate bonds, that can be an incentive for corporations to issue new bonds they wouldn't otherwise.

Plus, of course, the Fed can drive down the interest price, making more investment projects positive NPV and thus a go, increasing the velocity of "money", and economic activity, however you may define money.

Anonymous said...

What Musgrave said, except I would add that the Fed is going to lose on the round-trip, so there will eventually be some money-printing:

1. Fed announces QE, either openly or by "telegraphing'. Price of bonds rises from 100 to 110.
2. Fed buys bonds at 110.
3. Fed announces end of QE. Prices of bonds falls to 100.
4. Fed sells bonds at 100.

Bottom line, the Fed lost 10 on the round-trip and the rentier class picked up 10, as if they needed more money.

Roberto said...

I always taught my students that money has four functions: unit of account, medium of exchange, stand of deferred payment, and store of value. Whatever (SDR’s, cowrie shells, cigarettes, silver, bonds, artwork, salt, ‘bits’, demand deposits, land) performs these functions is money. What the one means by money depends on the context.

Warren Mosler said...

This and more at

Greg said...

Boy, I hope the comment thread in this post reaches the thousands!!

Calling all humans, come here and lets talk MONEY!!!

Anonymous said...

Can somebody discuss the issues that direct money creation (as in an unfunded budget deficit) can lead to once economic activity picks up? I always hear people dismissing these methods (and QE come to think of it) based on the difficulty of draining the resources back out at some futute date to prevent inflation.

jimmy wilson said...

It sounds pretty good

Fight Club Brad Pitt Red Jacket

sbobet 338a said...

This can be impressive. Scientists watch site’s video material and we are thunderstruck. We are curious about one of these things. Only appreciate your regular suggestion, and assess doing in this. Please keep enhancing. These are remarkably costly im. judi bola online indonesia