« Intermediate Macro Exercise: How to Construct a Maximally Contractionary Budget Deficit Reduction | Main
I wanted to offer some clarification on stories about all the money that the Federal Reserve is supposedly printing. It depends, I guess, on your definition of "money." And your definition of "printing."
When people talk about "printing money," your first thought might be that they're referring to green pieces of paper with pictures of dead presidents on them. The graph below plots the growth rate for currency in circulation over the last decade. I've calculated the growth rate over 2-year rather than 1-year intervals to smooth a little the impact of the abrupt downturn in money growth in 2008. Another reason to use 2-year rates is that when we're thinking about money growth rates as a potential inflation indicator, both economic theory and the empirical evidence suggest that it's better to average growth rates over longer intervals.
Currency in circulation has increased by 5.2% per year over the last two years, a bit below the average for the last decade. If you took a very simple-minded monetarist view of inflation (inflation = money growth minus real output growth), and expected (as many observers do) better than 3% real GDP growth for the next two years, you'd conclude that recent money growth rates are consistent with extremely low rates of inflation.
But if the Fed didn't print any money as part of QE2 and earlier asset purchases, how did it pay for the stuff it bought? The answer is that the Fed simply credited the accounts that banks that are members of the Federal Reserve System hold with the Fed. These electronic credits, or reserve balances, are what has exploded since 2008. The blue area in the graph below is the total currency in circulation, whose growth we have just seen has been pretty modest. The maroon area represents reserves.
Are reserves the same thing as money? An individual bank is entitled to convert those accounts into currency whenever it likes. Reserves are also used to effect transfers between banks. For example, if a check written by a customer of Bank A is deposited in the account of a customer in Bank B, the check is often cleared by debiting Bank A's account with the Fed and crediting Bank B's account. During the day, ownership of the reserves is passing back and forth between banks as a result of a number of different kinds of interbank transactions.
To understand how the receipt of new reserves influences a bank's behavior, the place to start is to ask whether the bank is willing to hold the reserves overnight. Prior to 2008, a bank could earn no interest on reserves, and could get some extra revenue by investing any excess reserves, for example, by lending the reserves overnight to another bank on the federal funds market. In that system, most banks would be actively monitoring reserve inflows and outflows in order to maximize profits. The overall level of excess reserves at the end of each day was pretty small (a tiny sliver in the above diagram), since nobody wanted to be stuck with idle reserves at the end of the day. When the Fed created new reserves in that system, the result was a series of new interbank transactions that eventually ended in the reserves being withdrawn as currency.
All that changed dramatically in the fall of 2008, because (1) the Fed started paying interest on excess reserves, and (2) banks earned practically no interest on safe overnight loans. In the current system, new reserves that the Fed creates just sit there on banks' accounts with the Fed. None of these banks have the slightest desire to make cash withdrawals from these accounts, and the Fed has no intention whatever of trying to print the dollar bills associated with these huge balances in deposits with the Fed.
Of course, the situation is not going to stay this way indefinitely. As business conditions pick up, the Fed is going to have to do two things. First, the Fed will have to sell off some of the assets it has acquired with those reserves. The purchaser of the asset will pay the Fed by sending instructions to debit its account with the Fed, causing the reserves to disappear with the same kind of keystroke that brought them into existence in the first place. Second, the Fed will have to raise the interest rate it pays on reserves to give banks an incentive to hold funds on account with the Fed overnight.
Doing this obviously involves some potentially tricky details. The Fed will have to begin this contraction at a time when the unemployment rate is still very high. And the volumes involved and lack of experience with this situation suggest great caution is called for in timing and operational details. Sober observers can and do worry about how well the Fed will be able to pull this off.
But that worry is very different from the popular impression by some that hyperinflation is just around the corner as a necessary consequence of all the money that the Fed has supposedly printed.
Posted by James Hamilton at February 16, 2011 07:04 AM
OT:
I took a crack at trying to estimate the timing of a next oil shock.
You can read the resulting article here:
http://www.aspousa.org/index.php/2011/02/an-oil-shock-in-2012/
Posted by: Steven Kopits at February 16, 2011 07:32 AM
Why don't the banks buy a higher yielding (longer dated) treasury note with this reserve money. It's highly liquid and pays alot more than 0.25%?
Posted by: dis737 at February 16, 2011 07:58 AM
James,
It is a good lesson explaining base money.
It would be interesting, as in fun if not particularly instructive, if you put a graph showing the actual measured amount of currency that has been printed but not destroyed. The number is many times the currency in circulation. Put it under the heading of "Where Did It All Go."
Another, more instructive, currency number you might publish is the amount of currency within the US border and that which is outside the world. The world still likes our little green pieces of paper.
Posted by: Ed Hanson at February 16, 2011 08:16 AM
what worries me is that the fed has paid top dollar for these assets and interest rates have moved unfavorably thus far. if the fed begins to sell, as you are suggesting, in order to withdraw liquidity while the treasury is running huge deficits, it will create a significant disbalance between buyers and sellers of treasuries (or agencies). this could have very disturbing impact on yields by first increasing the borrowing cost of the treasury and second by deepening the loss between what the fed has bought at and what it is selling at these assets.
do you really think that the fed can actually sell any assets before the budget deficit issue is rectified?
Posted by: baychev at February 16, 2011 08:37 AM
Do you have any evidence that the Federal Reserve credited money to the Bank's reserves as you claim?
If the Federal Reserve had actually created the approximately $1.5 trillion that was shifted into the reserve holdings, then this would presumably have appeared in the M2 figures as well, since they record the total of all bank deposits in the system. However there is no matching increase there - and it would stick out like a sore thumb with an increase on this magnitude.
What actually appears to have happened is that when the payment of interest on reserve accounts was announced Banks shifted existing money into reserve accounts at the Federal Reserve Banks. Of necessity electronically represented money must always be on deposit somewhere in the system. Whether they did this by reclassifying their depositor's accounts to attract a reserve requirement, or by simply putting their overnight funds in there, is hard to say - most probably both.
Further, in an Electronic Banking system the distinction being made here between physical currency and bank deposits is not very useful. Physical currency is currently printed on demand, and its growth is a reflection of the steady growth in the amount of money represented as on deposit in the banking system. The real question that needs to be addressed is why this increase has been occurring, independently of any QE measures, in almost all Basel regulated banking systems for over 10 years.
Posted by: cargocultist at February 16, 2011 09:07 AM
Will Bernanke be able to "pull this off" ?
My answer :
http://www.calculatedriskblog.com/2011/02/bernanke-hoocoodanode.html
Bernake quote for economic history : COMMISSIONER THOMPSON: Why did we -- and had we acted on them, might we have averted the disaster? MR. BERNANKE: Well, I don't know, I have to think about that.
Three years later and he still hasn't thought about it yet!
Posted by: Johannes at February 16, 2011 09:12 AM
JDH, Thank you for a very timely and relevant post. You write: "None of these banks have the slightest desire to make cash withdrawals from these accounts..." This indicates that the only use of these resources is loaning them out in the fed funds interbank market or keeping them on deposit at the Fed (incentivized by the interest on reserves). Is it possible that the opportunity costs of these excess reserves could change such that other alternatives become appealing to banks, leading to an outflow of reserves to other uses? The Fed claims that they could adequately respond to such a contingency (by raising the IOR or soaking up liquidity through reverse repos), but none of these actions have any historical precedence and leads to speculation that the excess reserves could lead to an increase in the money supply and inflation. Thanks for any comments.
Posted by: Rodney Chun at February 16, 2011 09:27 AM
JDH, You say "of course the situation is not going to stay this way indefinitely". I ask why not? What's the difference in banks holding reserves (which you recently called short term govt securities) or holding longer term govt securities? Excess reserves do not allow banks to make more loans. Some countries have zero reserve requirements. Banks in those countries make loans. If reserves were required for banks to make loans than banks in those countries would not be able to make loans. To make loans banks have capital requirements, credit requirements, and customer requirements, but reserves are not needed to make loans. Excess reserves will not lead to excess lending. So what's the big deal about excess reserves?
Posted by: markg at February 16, 2011 09:46 AM
James:
This is a modern version of the Law of Reflux in operation. Money that is not wanted in circulation will just return to its issuers. In the days of full-bodied coins, unwanted coins would be melted and would 'reflux' to bullion. Nowadays, unwanted paper money can reflux to reserves, and unwanted reserves can reflux to bonds.
Posted by: Mike Sproul at February 16, 2011 09:55 AM
cargocultist: You are confusing two completely different things: (1) deposits banks have in their accounts with the Fed, which are the reserve balances shown in the graph, and have never been included in M2; (2) deposits customers have in their accounts with their banks, which are included in M2.
dis737 and Rodney Chun: Absolutely, at some point longer term assets will be perceived as more attractive than short-term T-bills yielding essentially nothing, and that is the point at which short-term rates have to rise and the Fed will be forced to take the steps I discuss.
markg: The overnight interest rate on safe investments is not going to remain stuck at zero. When that changes, the Fed would have to increase the interest on reserves if it wanted to keep excess reserves at current levels.
Posted by: JDH at February 16, 2011 10:01 AM
How is what's happened different from the Fed basically stepping in and taking over the inter-bank loan market?
It appears that, instead of lending to each other on the overnight market, the banks are just lending to the Fed instead. My impression was that the motivation for these direct financing interventions came about not just to generate monetary stimulus but also to mitigate counterparty risk in the overnight lending market.
If you were to compare the value of total currency in circulation, reserve balances with Federal Reserve Banks, and the overnight loan balances between banks, would you still see an overall expansion in the monetary supply?
Posted by: Tudor at February 16, 2011 10:06 AM
"Second, the Fed will have to raise the interest rate it pays on reserves to give banks an incentive to hold funds on account with the Fed overnight." Balderdash!! The FED does not have to bribe banks to hold "excess" reserves, all it has to do is to RAISE the effective reserve requirement. If the argument is that some individual bank becomes encumbered because the FED decides to liquidate its overvalued mortgage security, I say: What kind of free ride do you want? I say the banking system is better off replacing badly managed over-leveraged institutions with healthy smaller banks. If the argument is that the credit (money) supply will contract if reserve ratios are increased. I say ...DUH... Isn't that the point? Isn't the problem that the level of debt in America exceeds the economy's ability to service it?
Read Full Article »