It's wrong to think of Fed transactions in terms of its profitability for the Fed

James Hamilton has a post stating that QE has added no interest rate risk for the Fed’s balance sheet, and that it has actually proven to be hugely profitable for the Fed. (H/T Mark Thoma) Where did the Fed get the money to buy all this stuff? The answer is, whenever somebody sold these items to the Fed, the Fed credited an account that the seller's bank maintains with the Fed in the form of new Federal Reserve deposits. At the moment, most of those new reserves are just sitting there at the end of each day on some bank's balance sheet. Reserve balances with Federal Reserve banks have gone from $9 billion in April 2007 to almost $1.5 trillion today.

The Fed is currently paying banks 0.25% interest on those reserves, and is collecting an average interest rate of 4% on its long-term securities. That netted the Fed a healthy profit of $80 billion in 2010, which it returned to the U.S. Treasury. In effect, the Fed is borrowing short and lending long, making a huge profit on the difference, and handing it back to the Treasury.

Prof. Hamilton’s analysis is wrong on several levels.
1. The Fed, which has the capability to credit money to accounts, i.e., print money from ‘thin air’ can never lose money, so it doesn’t make sense to justify its actions on the basis of it being profitable for the Fed.

2. In fact, if we were to take this line of thinking, and QE is such a good profit-making strategy for the Fed, then it should just keep on printing/crediting bank accounts for as much as needed to buy out all existing positive carry securities in the market, then pay correspondingly less interest (IOR) on the reserves such move creates for the banks. If the prevailing logic is to be believed, this takes away all market risks from the banks, puts them into the Fed’s balance sheet, and it provides banks with a high level of reserves with which “"to fund more lending"”, while at the same time piling on more profits for the Fed. What’s not to like?

3. But not all is hunky-dory, and the whole structure just manages to keep standing precisely because the Fed could print fiat money at will. Hence, it doesn’t make sense to justify the Fed’s moves on the basis of its profit-making effect on the Fed’s income statement. If you’re a private sector entity, which unlike the Fed, has revenue and cost constraints, doing exactly what the Fed is doing is a recipe for money-losing disaster. Where in the private sector will it make sense to buy out another party’s securities, at market cost, and then turn around and compensate that third part by paying him interest (IOR) on the money he ends up with from your buying? Where do you get off paying someone for value, then paying him again interest for the money you just gave him, then booking profits just because you paid him less than what you earned from the securities you bought from him? If you bought at market value, then you’ve already paid the value of all future interest to him, and you are just breaking even when you in turn receive it over time from the security. Any additional you pay on interest to the third party will already be a loss to you.

4. The risk of accounting loss for the Fed does not come from ever having to pay more in IOR than it’s earning from the securities it just bought (Why would that ever happen when it already paid the banks just compensation for the securities?) The risk of loss will come when the Fed decides to reverse its market buying, and start selling the securities back to the market. Because the Fed is a large entity undertaking a large market transaction, anything it does moves the market – against itself. When its buys, it adds demand to the securities, making it more expensive for itself when it buys. When it sells, it does the opposite. It removes a large block of demand from the market, causing a price decline when it sells to the market. What the market gains from this, the Fed will lose.

5. So will all this QE end up making a profit for the fed? No, it never did, and the fact it’s paying IOR normally should be a loss to the Fed. But it doesn’t because the Fed just prints the IOR it pays. And if it ever decides to reverse QE, money printer or not, it will incur an accounting loss. But don’t fret, the Fed is already on it.

Addendum: This is to address Detroit Dan in comments, about the Yves Smith post in the link - The Yves Smith post is a discussion of how the Fed plans to handle any interest rate/market loss it could incur if it decides to sell the hoard of Treasuries it bought via QE2.

Basically, if the selling results in a net loss to the Fed, it will just consider the current loss to be deductible from the amount of profit it sends to the Treasury the next time it has a profitable year. So during the year that it incurs the loss, it actually receives a "credit" from Treasury.

A comparable scenario in private business is if you have a business that incurs a loss this year, you just say "it's okay,I will in the meantime just consider the loss as a shareholder loan from me payable to me the next time the business is able to". This can go on indefinitely if you are like the Treasury, which can just debit-credit the losses until the Fed finally is able to pay again. But you, as a currency user, will have to fund the loss with actual money, while the Treasury, a currency issuer, doesn't.