His is a unique business, since he can force the market to lend him money—he simply buys what he wants and pays for it with his "bank reserve checkbook." By the end of next year, the Fed will own $1 trillion more bonds, and the banking system will have $1 trillion more reserves, whether it wants them or not. Bernanke can also dictate the rate at which he borrows money; for the foreseeable future that will be the rate the Fed decides to pay on reserve balances held at the Fed, currently 0.25%. Those who end up with the reserves will have essentially lent the Fed money on the Fed's terms.
To be more specific: Next year, Bernanke plans to make net purchases of $540 billion of longer-term Treasuries, and $480 billion of MBS. He will fund those purchases by issuing $1.02 trillion of newly-minted bank reserves. In effect, the Fed will be swapping reserves (which are functionally equivalent to 3-mo. T-bills, the paragon of risk-free assets, but which currently pay a slightly higher rate of interest) for bonds. Since money and bank reserves are fungible, Bernanke's planned purchases should effectively cover Treasury's deficit next year, which, perhaps not coincidentally, looks to be about $1 trillion.
It's important to note here that when the Fed issues $1 trillion of bank reserves, it is NOT "printing money." That's because bank reserves are not cash and they can't be spent anywhere: like pajamas, they are only for use "in house," since they are always kept at the Fed. Bank reserves do have a unique feature, of course, that other short-term assets don't: they can be used by banks to create new money, and in fact, acquiring more reserves is the only way that banks can increase their lending, because banks need reserves to back their deposits. Since banks now hold $1.6 trillion of reserves, of which only $0.1 trillion is required to back current deposits, banks already have an almost unlimited ability to make new loans and thereby expand the money supply. A year from now they will have an even more unlimited ability to do so.
That banks haven't yet engaged in a massive expansion of lending activity and the money supply is a testament only to the risk-averse nature of bank management and the risk-averse nature of the public, which now holds $6.5 trillion of bank savings deposits (up 64% in the past four years) paying almost nothing. As the above chart shows, in recent years the M2 measure of money supply has grown only slightly faster than its long-term average.
To put it another way: The Fed's massive provision of reserves to the banking system has not resulted in an equally large increase in inflation because the world's demand for money (cash, bank deposits, and cash equivalents like bank reserves and T-bills) has been very strong. Banks, in short, have been content to sit on $1.5 trillion of "excess" reserves because they worry that making more loans and increasing deposits might be a lot riskier.
The rationale for hedge funds is to exploit arbitrage opportunities, buying one thing and selling or borrowing another. Even small differences in prices can become lucrative, thanks to the use of lots of leverage. If done successfully, arbitrage can contribute to market efficiency, which in turn can contribute to the health of an economy. Whether the Fed will accomplish the same thing with QE4, however, is an open question. Will banks lend a lot more next year, even though they have an essentially unlimited capacity to lend today? Will increased bank lending fuel genuine economic growth, or will it just fuel more speculation? No one knows. We are in uncharted waters; what the Fed is doing today has never been done before.
When faced with issues of daunting complexity and with little or no guidance from the past, one can only begin by trying to reduce things to their simplest form. Here's what I think is a simplified description of what the Fed is planning: Next year the Fed will be purchasing a total of $1 trillion of 10-yr Treasuries and current coupon MBS. 10-yr Treasuries currently yield 1.75%, and current coupon MBS about 2.25%, so the Fed will earn roughly 2.0% on its purchases, while paying out 0.25% on the reserves it creates to buy those bonds, for a net spread of 1.75%. By the end of next year, the Fed will be raking in $17.5 billion per year in profits on their $1 trillion swap, and that will make the Fed the envy of all other hedge fund managers.
These profits, of course, are automatically remitted by the Fed to Treasury. Happily for taxpayers, those profits will completely offset Treasury's cost of borrowing, at least for the next several years. Here's the math, also in simplified form: First, let's assume that Treasury is funding its deficit with 7-yr Treasuries (that's a decent approximation, since last year they told us that they were going to lengthen the average maturity of outstanding Treasuries, which at the time was about six years). The yield on 7-yr Treasuries is currently about 1.25%, so Treasury will pay 1.25% on $1 trillion, and receive back from the Fed 1.75%, leaving a profit of about 0.5%, or $5 billion. Bottom line, we will all benefit from next year's deficit financing! (Note that the key to the profit is the Fed's decision to buy lots of MBS, which yield more than Treasuries of similar maturity.)
A real-world hedge fund attempting to do the same thing would run up against the reality of mark-to-market accounting rules. If interest rates on the bonds it buys rise, the mark-to-market losses on the bonds could easily wipe out the interest it's receiving, threaten margin calls and ultimately result in insolvency. For example, a 1 percentage point rise in the yield on 7-yr Treasuries would result in a 6.7% decline in their price, thereby wiping out over 5 years' worth of coupon payments. Mortgage-backed securities could fall in price by even more. A hedge fund would also be exposed to the risk that its borrowing costs could rise, thus narrowing or even eliminating the net interest spread it's earning.
Happily, Bernanke doesn't have to worry about any of this, since he doesn't have to mark his bonds to market, and he can keep his borrowing costs below the current yield on his portfolio for at least the next 2 or 3 years, given the FOMC's recent guidance (i.e., it won't start tightening until the unemployment rate falls to 6.5%, short-term inflation expectations exceed 2.5%, and/or long-term inflation expectations become unanchored). And of course, the Fed can always make the interest payments on its borrowings because its "bank reserve checkbook" is effectively bottomless.
If this all sounds too good to be true, it is. The Fed may not face the risks that a typical hedge fund does, but that doesn't mean the Fed is not taking on a huge amount of risk at taxpayers' and citizens' expense. Although the Fed need never face insolvency, if mark to market losses got really bad, they could lose their credibility and with that the value of the dollar could be seriously at risk. The Fed's losses might become direct obligations of Treasury, or they might be inflicted on taxpayers and citizens via the sinister "inflation tax." The Fed could eventually repay its borrowings with devalued dollars, leaving the rest of us with deflated balance sheets and deflated incomes. Meanwhile, by allowing Treasury to borrow trillions at no cost, the Fed is acting as an obstacle to badly needed deficit reduction.
Although it may seem paradoxical, the biggest risk we all face as a result of the Fed's unprecedented experiment in quantitative easing is the return of confidence and the decline of risk aversion. If there comes a time when banks no longer want to hold trillions of dollars worth of excess bank reserves for whatever reason (e.g., the interest rate the Fed is paying is no longer attractive, or the banks feel comfortable using their reserves to ramp up lending, or the public no longer wants to keep many of trillions of dollars in bank savings deposits), that is when things will get "interesting."
More confidence would mean less demand for cash and cash equivalents, and that in turn would mean that a virtual flood of money could try to exit banks (e.g., as people withdraw their savings deposits, and/or borrow more from their banks). If the public attempted to shift trillions in cash into housing, stocks, gold, or other currencies, the consequences would likely be seen in sharply rising prices and higher inflation. Moreover, higher inflation would almost certainly lead to higher interest rates, which in turn would exacerbate the Fed’s mark to market problem and possibly accelerate the whole process. And of course, higher interest rates will result in significantly higher borrowing costs to Treasury, although this will be mitigated to some extent by Treasury's efforts to extend the average maturity of its borrowings.
The Fed reasons that it could deal with declining risk aversion by selling bonds (i.e., reducing bank reserves), not reinvesting principal, and by raising the rate it pays on bank reserves. But it’s not hard to see how things could get out of control: higher rates on bank reserves would likely accelerate the rise in market yields and the mark to market losses on the Fed’s bond holdings, at the same time as its spread eroded. In the meantime, the more bank reserves the Fed creates, the harder it will be to avoid an unhappy outcome.
It’s ironic that the Fed is trying, with QE4, to accomplish the very thing that could be its own undoing. Trying, that is, to encourage more confidence, more lending, more borrowing, more investment, and higher prices for risk assets.
It’s no wonder that the market remains so risk-averse, since this is hardly a comforting position we're in. For now, that is probably a good thing. But in the wake of the election results and the Fed's latest decision, I am less optimistic today than I have been for several years.