Friday, December 10, 2010

There Is No Such Things As Monetary Policy

In my last post, I argued that bank reserves, if they pay interest, are essentially a form of government debt. They’re issued by a different institution and have a different maturity than most government debt, but in their essential nature, they’re just a special case of government debt.

In that post, I was agnostic about whether bank reserves should be considered “money.” If bank reserves are government debt, then, to be consistent, we should either consider all government debt to be money or consider bank reserves to be something other than money. In particular, Ben Bernanke’s use of the phrase “printing money” is consistent if you consider all government debt to be money, in which case QE2 only exchanges zero-maturity money for high-maturity money whereas QE1 exchanges money for private sector assets. If you take the second option, however, then neither QE1 nor QE2 was “printing money.”

I want to explore that second option, which seems pretty reasonable in terms of the way I learned liquidity preference theory in school. In the simplest version of liquidity preference theory, money (1) pays no interest and (2) is controlled by policymakers. Obviously bank reserves no longer meet the first criterion, and bank deposits (as we learned painfully during the policy experiment of the early 1980’s) do not meet the second criterion. I’m increasingly coming to believe that the most reasonable definition of “money” is simply “cash” (negotiable central bank notes plus coins).

But the Fed does not attempt to control the supply of cash. Banks can withdraw reserves from the Fed, or deposit them, whenever they choose, and the Fed will normally compensate for aggregate net withdrawals or deposits by adjusting the total level of reserves to bring it back to where it was before the net withdrawals or deposits took place. Moreover, the amount of cash outstanding, as determined by the demand for cash, is not a significant factor in the setting of Fed policy. If “money” simply means cash, then the Fed’s “monetary policy” is unlimited accommodation. And moreover, nobody cares, or should care, about how this policy is going. For practical purposes, there is no such thing as monetary policy.

The Fed does do something active, and they call it monetary policy. But what is it really? The Fed does several things that usually come under the heading of monetary policy, but they can all be reasonably classified as something else. For one thing, it manages the level of bank reserves through open market operations. But this is really government finance policy, not monetary policy: it consists of substituting zero-maturity government debt (bank reserves) for higher maturity government debt (Treasury securities), or vice versa.

The Fed also sets the interest rate on reserves. This is also government finance policy. Suppose that, instead of auctioning off fixed quantities of securities, the Treasury were to sell as much as it could at a given interest rate and leave any excess cash in a vault. If we’re talking about, say, one-month T-bills, which have only slightly higher maturity than bank reserves, then the effect would be essentially the same as when the Fed sets the interest rate on reserves. (In particular, imagine that, starting from today’s near-zero interest rates, the Treasury were to start selling T-bills with a fixed 1% yield. That would be just like an increase on the interest rate on reserves, except that others besides banks could participate. That difference is really irrelevant except for the arbitrage opportunity provided to banks when the IOR rate exceeds the T-bill rate. If the IOR rate were to rise, banks – being in competition with one another for funds – would have an incentive to let others participate indirectly.)

Granted, this doesn’t work in reverse: the Treasury needs to borrow a certain minimum amount to finance its expenditures, so it can’t set an arbitrarily low interest rate. But it doesn’t work in reverse for bank reserves either. If the Fed sets the IOR rate far below the T-bill rate, it becomes irrelevant: banks will simply hold, for emergency purposes, a tiny amount of reserves over and above what is required, and the rest of their reserve assets they will hold as T-bills, which can easily be sold or repoed for federal funds if necessary.

The Fed also sets the reserve requirement, thereby compelling a demand for reserves, but this is not monetary policy either; it is regulatory policy. There is a regulation that requires banks to hold a certain quantity (set by the Fed) of zero-maturity government debt (reserves). There are other regulations that effectively require banks to hold certain quantities of more broadly defined safe assets to satisfy capital requirements. All these regulations create an increased demand for government debt; it’s only the maturity of the debt that is different. To the extent that such policies are used for macroeconomic demand management, they are essentially a form of taxation: the Fed can pay a lower rate of IOR than it otherwise would because it is compelling banks to hold reserves instead of using the money for profitable activities; it could achieve the same effect by raising IOR until those activities are no longer profitable (in risk-adjusted opportunity cost terms) and taxing the banks an amount equal to the difference in the interest they pay. To the extent that regulatory policy is used for macroeconomic demand management, it is just a form of fiscal policy.

Finally, the Fed lends (or purchases the assets from those who have lent) to the private sector, via its discount window, via various emergency programs in times of crisis, and in particular via QE1. But this isn’t monetary policy; it’s fiscal policy. It’s no different than what the TARP did: an exchange of government debt for private debt. Anything the Fed does through such programs could also be done by the Treasury, which could issue T-bills to obtain funding, and it would surely be considered fiscal policy. The Fed issues bank reserves instead of T-bills: a slight difference of maturity, not a fundamental difference of substance.

So there you have it: there is no such thing as monetary policy. There is “central bank directed stabilization policy,” and, for convenience, you can refer to that as monetary policy if you want. If so, recognize that you are using the term loosely, and let’s not get into arguments about whether some particular Fed action is “really” monetary policy. None of it is really monetary policy.

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.


Anonymous said...

I think Bernanke was using a technical detail to score a rhetorical point in this particular interview. That detail is consistent with your new definition of money. Based on that definition, the monetary policy is to provide money on demand, in exchange for bank reserves.

Regarding your evolving insights on reserves, etc., which are sound, you should really have a look at the blogs of Bill Mitchell, Warren Mosler, and “neweconomicperspectives” from Kansas City. You’re stumbling upon stuff they revealed a long time ago. See "billyblog" archives; anything with “bank reserves” in the title.

Nick Rowe said...

It's very good to see you back blogging Andy. I missed your posts. Always both deep and simple.

Gonna think this one through, then see if I can totally demolish you, for daring to post such an inflammatory title!

Torgeir Høien said...

Andy, sorry about posting before proof reading. Please delete my earlier version.

You’re absolutely right, there is no monetary policy. QE1 was credit policy and QE2 is debt management policy. These initiatives could, and should, have been done by the Treasury, as they are risky and have fiscal implications. The Treasury, of course, did its own version of QE1 (TARP), and it does debt management policy, be it QE or QT (Quantitative Tightening). Since late 2008 the Treasury has done QT, as the average duration of the federal debt has increased.

What the Federal Reserve does is interest rate policy. With the monetary unit – the dollar – defined as an arbitrary denomination the public sectors liabilities, someone in the public sector, usually the central bank, has to fix a nominal interest rate. Central banks typically do this by fixing the overnight interest rate on banks’ central bank deposits. Since October 2008, when Congress gave the Federal Reserve’s Board of Governors authority to pay interest on deposits, this has been done administratively.

Interest rate policy too has fiscal implications. Normally these are small. But imagine a situation where the Federal Reserve has bought all the public debt, and the public sector is financed solely through overnight central bank debt. Then interest rate policy has a huge impact on public sector interest expenses, and monetary tightening is expansionary if the Treasury doesn’t neutralize the increase in interest expenses by cutting other expenditures or raising taxes.

Central Bank independence means 1) that the Treasury delegates interest rate fixing to the central bank and 2) that the Treasury adjust fiscal policy passively to neutralize the impact interest rate policy has on public finances. Time will show whether the Federal Reserve can stay independent from the Treasury now that it has become a Treasury of its own.

OGT said...

I look forward to Nick Rowe's reply. I doubt Quasi-Central Bank Stabilization Policy-ist is a title many economists would want to flock towards.

The comparison of money and credit feels vaguely PK, though.

Greg said...

Seems to me Andy, with your last two posts, you are coming very close to the MMTers view of the world.

Just sayin'

Bill Woolsey said...

The problem with this analysis is at the beginning. Paying interest on reserve deposits at banks makes them into a type of government debt. And then, all government debt is money and off to the confusion.

Your training was in error. The version of monetary theory where money has no interest and is controlled by policy makers is--well, not necessarily so.

For some years now, most money has paid interest. There is no reason why all money might not pay interest.

Money is the media of exchange. It is a set of assets generally accepted in exchange.

If the central bank is owned and operated by the government, then all of its liabilities, whether paying interest or not, is a type of government debt. Hand to hand currency is a type of governemnt debt that has a zero nominal interest rate. And, it serves as media of exchange.

Reserve balances at the Fed are another type of government debt, that currently pay interest, and also serve as media of exchange.

Perhaps there are some markets where T-bills serve as media of exchange, but not all government bonds do.

Deposits at private banks often serve as media of exchange. They pay interest. But they are not government debt.

If the central bank is private, or it's legal status is so independent that calling its liabilities "government debt" is like aggregating the debts of South Carolina and Georgia, then it should be treated more like a private bank. Its deposits and currency are its liabilities and not the debt of the U.S. Treasury or New Jersey or Los Angeles.

Under that scheme, the central bank holds a variety of assets, much of which is government debt (or the debt of other governmental entities.)

Still, the currency it issues and the deposits it creates serve as media of exchange.

David Beckworth said...


I agree with Bill. What makes money unique is its medium of exchange role. Any asset that takes on a true medium of exchange role has no market of its own. Thus, shocks to demand or supply for the medium of exchange cause disruptions in other markets. Monetary policy is about responding to these medium of exchange shocks such that disruptions to other markets are minimized. This is far different from fiscal policy.

If you are interested, Leland Yeager spells out it more fully here:

Andy Harless said...


"Any asset that takes on a true medium of exchange role has no market of its own."

I don't think this is true. There are multiple media of exchange, and there are markets where they trade against one another.

"Monetary policy is about responding to these medium of exchange shocks such that disruptions to other markets are minimized."

But the same could be said for fiscal policy, coludn't it? If people are hoarding the medium of exchange, one way to fix it is for the government to spend more of it to make up for lack of spending by the private sector. Then where do you draw the line between fiscal and monetary policy?


In the US, it seems to me, the Fed is a wholly owned but independently operated subsidiary of the federal government. For financial purposes, it makes sense to consolidate their balance sheets. Or else the Fed should appear as a separate entity on the government's balance sheet, in which case bank reserves (and perhaps cash) would be like agency debt.

And base money bears little if any relationship to the total supply of media of exchange. Any asset that can be repoed is effectively a medium of exchange, as is the personal credit of an individual that makes purchases on credit (as most people do today, using plastic). If monetary policy is about manipulating the medium of exchange, then conventional monetary policy is only a tiny part of it. Instead of saying, "There is no such thing as monetary policy," perhaps I should be saying, "There is no such thing as something that is not monetary policy."

David Beckworth said...


I should be more precise. A fiat medium of exchange has no meaningful market of its own in terms of monetary disequilibria. For example, assume there is a sudden fear-driven spike in money demand and the Fed does not accommodate it with an increase in the money supply. There is no explicit market for the medium of exchange where this disequilibrium can be worked out. Thus, it has to be worked out through other markets. And if prices are sticky in these other markets then there will be a real economic disruption as well.

Yes, there is the forex market, but this is just a market where one can buy an asset that serves as a medium of exchange somewhere else. It provides a market for trading one currency for another, but not for the demand and supply of single currency--and that is what is needed to work out monetary disequilibria.

Regarding the efficacy of fiscal policy versus monetary policy, you have a point. Fiscal policy can offset drops in velocity. But it is still limited relative to monetary policy for two reasons.

First, while in principle fiscal policy can offset changes in velocity it has less ability to do so with changes the money supply. It could influence inside (i.e. broad) money creation, but less so with outside (i.e. base) money creation. Yes, IOR reserves weakens even that last claim, but at a minimum monetary policy alone shapes currency in circulation.

Second, monetary policy can always trump whatever fiscal policy attempts to do. Monetary policy can act more quickly and has a larger arsenal with which to work.

Anonymous said...

Thank you for this very persuasive piece. I think it is important to note that when the government spends it creates reserves and when the government collects taxes it destroys reserves. No other entity can do this without taking away or putting back in some other asset -- not even the Fed.

As you mention, even in the event of a 'shock to the demand or supply of the medium of exchange', the government can simply spend or tax by the amount necessary to address the 'shock'. TARP is the perfect example of this, an example which also shows how flexible and precise fiscal operations can be.

Imagine a government that spends directly, pays no interest and issues no debt. If there is a sudden demand for the media of exchange, it spends more, and if the media of exchange starts to devalue, it taxes more. None of this would be considered monetary policy, but I don't see how this government would have any less flexibility or power than a government with a central bank.

Alice Cook said...

Thanks for the comment on my clouded outlook blog, I have just posted a reply.