At the Federal Reserve’s July 13, 2011 hearings before the House Financial Services Committee, Representative Al Green cited evidence that the Fed made a profit on its QE 1 and QE 2 asset acquisitions and that the transfer of those funds back to the Treasury has helped reduce the deficit. In response, Chairman Bernanke stated that the Fed would make a profit and the proceeds remitted to the Treasury would help reduce the deficit.
The essential argument is that the Fed has earned interest income on its large holdings of securities. After deducting expenses and making the required contributions to surplus and capital, the remainder is remitted to the Treasury as “profit” and is scored by the Treasury as revenue. The sums are huge; and last year, for example, according to the Fed’s 2010 annual report, it paid $79.268 billion to the Treasury.
From the Fed’s perspective, this transfer of funds may look like a remittance of “profits,” but when one looks at the government’s consolidated balance sheet and income statement, these are not profits from the taxpayer’s perspective. In fact the Fed cannot make a profit for the taxpayer related to its asset acquisition activities, whether as part of the bailout or in its normal course of business. To understand why, it is necessary to engage in what some readers will regard as a mind-numbing discussion of Treasury and Federal Reserve transactions and accounting.
The easiest way to understand the relationship between the Treasury and the Fed is to follow a series of transactions as the Treasury borrows from the public and makes expenditures. Suppose the Treasury decides that Secretary Geithner needs a new limo. To generate the necessary funds for the purchase, the Treasury first borrows from the public rather than raising taxes. It does so by selling a T-bill in the market. The public pays for the T-bill by writing a check that is deposited by the Treasury with the Fed. The Fed settles the transaction by transferring ownership of a reserve deposit at the Fed from the check writer’s bank to the Treasury’s account.
There is no impact of this transaction on the net worth of the private sector at this point since, all that has happened is that the purchaser of the T-bill exchanged one asset (a bank deposit) for the T-bill. To complete the transaction the Treasury writes a check on its account at the Fed to pay for the limo. Note that the Treasury’s balance sheet has increased, since the government now owns a limo, with the offsetting liability being the outstanding T-bill. When the check is deposited, the car dealer’s bank’s reserve account at the Fed is increased. However, the private sector’s net worth remains unchanged, since the car dealer has exchanged ownership of the limo for a check drawn on the Fed.
Let us now bring the Fed into play. The Fed engages in a quantitative easing transaction, as it did during QE 1 and QE 2, by purchasing from the public the T-bill the Treasury previously issued. It pays for the T-bill by writing a check that, when deposited, results in an increase in the seller’s bank’s reserve account. At this point, the Fed’s assets and liabilities have both gone up by the amount of the cost of the limo.
The Fed now holds a T-bill as an asset, which exactly equals the Treasury’s outstanding liability. But the Fed, too, is part of the government, which means that the government owes itself money. One option would be for the Fed and Treasury to mutually agree to cancel the T-bill, since the government now owns a claim on itself. From a consolidated balance sheet perspective (which is what the taxpayer cares about), the Treasury owns a limo, the T-bill is a wash, and the Fed has now increased the amount of money in the financial system by the amount of the cost of the limo. Put another way, the Fed printed money, and indirectly monetized the Treasury’s debt issuance. The Fed could simply have purchased the T-bill directly from the Treasury but for the fact that it is legally prohibited from doing so.
While from the taxpayer’s perspective the Treasury debt has been extinguished, this is not the case from the perspective of either the Fed or Treasury individually. Treasury pays the Fed interest on its holdings of Treasury debt. The Fed in turns uses the proceeds to cover its operating expenses, and then transfers the remainder back to the Treasury (we are ignoring, here, any other revenue generating activities the Fed may have). Under current accounting conventions the Treasury is permitted to treat the receipt as revenue for budget and deficit calculation purposes. This is why Chairman Bernanke and others state that the Fed’s transfer has helped to reduce the deficit.
But this doesn’t make any sense. From the taxpayers’ perspective, the government paid interest to itself and in the process magically converted an expense into revenue. This is financial alchemy. Note, too, that in every case the amount returned to the Treasury by the Fed has to be less than what the Fed received as an interest payment. This is because the Fed first takes out its operating expenses, which have grown significantly, because it now is paying interest on reserves. For example, in 2011 the Fed received $82.690 billion in interest from government entities but only returned $ 79.268 billion to the Treasury.
Instead of paying interest and then remitting the residual, an equivalent transaction would be for the Treasury and Fed to agree to only transfer the net difference between interest owed and interest to be returned. This is exactly the kind of netting transaction that occurs in making payments on interest-rate swaps. This would result in a net payment by the Treasury to the Fed to cover its costs of operations and transfers to reserves. In every case, there is a net cost (but no net revenue) to the Treasury (and the taxpayer) because of the need to cover the Fed’s costs of operations and transfer to capital.
The cost to the taxpayer of operating the Fed can only come from either taxes or additional borrowings. In either case the taxpayer is ultimately responsible. But there is also another important potential cost when it comes to unwinding the money-creating process associated with QE 1 and QE 2 (or the limo purchase).
The process of printing money creates the threat of future inflation. If the Fed does nothing, then that future inflation can wipe out real wealth and the purchasing power of the taxpayer (which is a future cost of quantitative easing). Of course, the Fed can’t let this happen. The canceling out of the Treasury debt that I suggested above is also not really a feasible option, because the Fed would have a liability with no offsetting asset and the Treasury would have a limo but no offsetting liability (the T-bill). What if instead of cancelling the T-bill, the Treasury gave the Fed a non-tradable, non-interesting-bearing Treasury security in return for the T-bill? Their balance sheets would balance and there would be no effect on the government’s consolidated balance sheet, but now the Fed would have no revenue to cover its operating expenses and would have to get government appropriations to cover them. But the essence of the Fed’s independence is its ability to set its own budget and not be subject to the political appropriations process. This situation also makes it clear, again, why there is no Fed profit, only an expense.
Without tradable securities, the Fed would no longer be able to conduct traditional open-market operations. Instead, the Fed would have to rely upon either increasing reserve requirements or paying interest on reserves to sterilize the high-powered money it had injected into the system. The only way liquidity could be withdrawn would be for the Treasury to pay off the debt it had given the Fed. And this could only be done by issuing additional debt into the marketplace (presumably at higher and higher interest rates) or increasing taxes, both of which again represent costs to taxpayers. The more the Fed relies upon the payment of interest on reserves, the higher the cost to the taxpayer, since the expense would have to be covered through appropriations. The better option would be to increase reserve requirements and only gradually relax them as the economy grew to accommodate the additional expansion in the money supply, but this would only reduce the losses, not eliminate them.
Under the current situation the Fed has the options of engaging in open-market operations, selling assets, or letting securities run off as the need to withdraw liquidity presents itself. However, any of these options will be associated with higher interest rates, and selling assets will surely be accompanied by capital losses. These all represent costs to the taxpayer, because they will be covered out of lower remittances to the Treasury.
One final point. Lest one doubt that the Fed is part of the government or that the Treasury is the ultimate backstop to the Fed, Federal Reserve Notes, which are Federal Reserve liabilities, are deemed obligations of the US government under Title 12 (Chapter 3, Subchapter XII, Section 411) of the US Code. Furthermore, there was a recent agreement between the Treasury and Fed on how the Fed will account for its transfers to the Treasury. Part of that agreement provides that if there are insufficient funds to equate the Fed’s capital and surplus, then a deferred asset can now be created, the negative amount of which represents a claim on future revenues that will be extinguished before further transfers are made to the Treasury. This means that the Treasury is backstopping the Fed by providing implicit capital.
The bottom line is that it is conceptually improper to consider the net revenue of the Fed as profit or to treat the return of excess revenues as income to the Treasury. In every dimension, there is a net cost to the taxpayer.
Robert Eisenbeis, Ph.D.
Vice Chairman & Chief Monetary Economist
Email | Bio
