We've received numerous e-mails regarding the comment (here) that the Fed (or any other central bank for that matter) finances securities purchases with reserves. It's unfortunate that the internet is full of misinformation, propagated by both bloggers and the mass media. The Fed's operations are not a mystery - it's just basic accounting. And the fundamentals of accounting tell us that if you increase your assets by purchasing something, your liabilities increase as well. The balance sheet "has to balance".
When the Fed buys a security, any of the following could be taking place:
1. The Fed sells another security in the same amount (such as in Operation Twist).
2. The Fed can lend that security via repo. In this situation the increase in assets (security purchase) corresponds to increase in liability (the Fed borrows cash against the bond).
3. The Fed can accept time deposits (now it owes money on the deposit - thus increases its liability).
4. The Fed can in effect use the proceeds from the repayment of various emergency facilities to cover the purchase. This was the case during part of QE1 (see discussion from 2009).
5. The Fed can increase bank reserves (by simply crediting the seller's reserve account). Remember that reserves are liabilities on the Fed's balance sheet.
These are all different ways the Fed can finance securities purchases. Only number 5 represents outright quantitative easing. Unless 1-4 are involved, reserves are used to fund balance sheet expansion.
There are some silly notions about what banks can and can not do with their reserve balances. Keep in mind that cash is fungible. A bank can buy securities or loans from another bank and get rid of its excess reserves - thus converting reserves into other assets. Of course then the seller bank will be stuck with these excess reserves. Only the Fed can change the total reserve balance in the banking system as a whole by buying and selling securities or by borrowing and lending money.
From our sponsor: