As the Federal Reserve has continued to buy bonds to aid the economic recovery, critics of its actions, and even some supporters, have grown increasingly concerned about what comes after all of this “quantitative easing”: How will the Fed “unwind” its balance sheet — that is, sell off the bonds it has purchased — without harming the economy?
The short answer: Don’t worry about it. The Fed can reverse its actions without wreaking economic damage, especially if it does it at the same time as it announces that it intends to keep nominal income growing at a stable rate.
Nobody disputes that the Fed will at some point need to reduce its asset holdings. They have grown so large in the first place because the financial crisis accompanied (and, in our view, to a very large extent resulted from) a sharp increase in the public’s demand for money balances: for the safety, that is, of cash and its near-equivalents. The Federal Reserve increased the supply of money in response to this increase in demand. It did so by purchasing Treasury and agency securities from the public, thus adding to the money held in bank accounts. The increase in supply, however, was insufficient to keep up with the increase in demand, especially in 2008 and 2009, which is why the crisis was so severe.
The demand for money balances and safe assets is still very elevated, albeit down from its crisis peak. If the economy enters a robust recovery, that demand should fall, as it usually does when there are attractive alternatives to just holding on to money. At some point, for example, banks will want to start investing more aggressively the $1.7 trillion in excess reserves — dollars they hold beyond what they are legally required to hold — that they have accumulated since the crisis began. A strong recovery would raise the demand for credit and provide just such an investment opportunity for banks in the form of higher-yielding loans. This increased lending would increase the money supply and, left unchecked, would cause a rapid rise in inflation. Recall the old explanation of what causes inflation: too much money chasing too few goods.
Fed officials say they plan to use two different methods to prevent such a destabilizing surge in the money supply. First, the Fed will stop reinvesting principal payments it receives from its maturing Treasury and agency securities. The dollars the Fed takes in, that is, will no longer circulate. Since the Fed won’t be buying more securities at this point, its balance sheet will shrink. The Fed learned the hard way how effective this approach can be: By not reinvesting the proceeds when its agency securities matured between mid 2010 and late 2011, it passively drained about $662 billion from the economy. It decided to reinvest the payments so as not to tighten monetary policy inadvertently. In the future, it will deliberately tighten money by again ceasing to reinvest.
The second method the Fed will deploy to shrink its balance sheet is to sell off its agency securities. According to the minutes from the June 2011 Fed meeting, the Fed would most likely do this over three to five years.
Along with taking these two measures, the Fed would have to adjust its “forward guidance” on interest rates — that is, raise its projections of where it expects interest rates to go — and gradually raise its target federal-funds rate. A recent Fed study found that, under reasonable scenarios, this unwinding process would take about five years and would cause no disruptions to economic activity.
One source of worry about this process is the fear that it will be expensive for the Treasury. As interest rates go up, some of the Fed’s securities will decline in value and will have to be sold at a loss. Also, the Fed has been paying banks interest on excess reserves, and higher interest rates will increase the size of these payments. The Fed has been sending money to the Treasury, but these developments would reduce that flow. In a worst-case scenario, the Fed could experience an operating loss and require a taxpayer bailout.
#page# The worriers are overlooking some important facts. First: While the Fed will send fewer funds to the Treasury during the winding down of its balance sheet, it will have sent, over the period including both the expansion and the unwinding, higher payments to the Treasury than normal. Prior to the crisis, the Fed was earning about $25 billion a year on average. Since then, it has earned as much as $90 billion, and it is expected to earn $40 billion on average over the 2009–25 period. We should not distort monetary policy to keep payments close to an abnormal peak.
Second, an operating loss need not mean a bailout. Imagine, for example, that the Fed were to take a large capital loss on its holdings and could not sell enough securities to rein in a rapidly expanding money supply. It could still tighten money by raising the interest rate paid on excess reserves, and fund the higher payments by tapping into its future earnings. It could do this by creating more dollars to pay the banks and then offsetting this transaction in the future by sending less of its earnings to the Treasury.
Third, these concerns miss the forest for the trees. These potential balance-sheet problems will emerge only if there is a robust economic recovery. We would be fortunate to have these problems! Moreover, a robust recovery would mean much more tax revenue for the Treasury. Even the best year of Fed earnings pales in comparison with the annual trillion-dollar deficits the weak economy has brought. Surely, no one would argue that we should keep the economy weak so that the Fed can generate extra revenue for the federal government.
The impact on the Treasury isn’t the only widespread concern about the Fed’s future undoing of quantitative easing. One popular line of argument holds that the economy has been “artificially” boosted by the Fed, which has been inflating a bubble as it inflated one in the pre-crisis years. Based on this premise, some people argue that the Fed now faces a no-win choice between popping the bubble and inflating it further. So either the Fed will not unwind its balance sheet, and we will get galloping inflation, or it will and the economy will crash.
These concerns, too, are overblown. The claim that the Fed has been inflating a bubble, for example, is based on the idea that its monetary policy has kept interest rates below their natural levels. That’s not true. Low interest rates are almost entirely the result of a weak economy, not the Fed’s inadequate attempts to loosen money.
As we’ve argued in these pages before, the key to distinguishing between loose and tight monetary policy is what’s happening to nominal income: that is, to the size of the economy measured in dollar terms, with no adjustment for inflation. If the growth of nominal income is accelerating, then monetary policy is loosening; and if growth decelerates, it’s tightening. The right policy aims for steady growth, which happens when the supply of money rises and falls with demand.
By this measure, Fed policy was loose in the years before the crash, leading to interest rates below the natural level and rising household debt. But Fed policy has since then been very tight — first disastrously tight in 2008–09, then only damagingly tight.
One reason to keep nominal-income growth steady is that most debts, such as mortgages, are contracted in nominal terms. An unexpected slowdown in its growth, as we have experienced over the last five years, makes the burden of repaying those debts heavier. The fact that households are nonetheless continuing to deleverage, rather than to add to their net borrowings, suggests that this is not a bubble economy.
#page# The claim that the Fed cannot be trusted to unwind its balance sheet — that it will let inflation go out of control — ignores its actual record over the last five years. Inflation has consistently come in below the Fed’s target and unemployment above it: The Fed has erred, that is, on the side of tightness. The last five years have seen a lower average inflation rate than any five-year stretch since the mid 1960s. Yet all of the political pressure on the Fed from Congress has been directed at getting it to tighten more.
Finally, we have a tool that we lacked the last time we experienced high inflation: a market-based indicator that will warn us that it is on the way. The Treasury now issues bonds indexed for inflation as well as unindexed ones, and the spread shows market expectations of inflation over the duration of the bond. If that spread rises a lot, the Fed will face even more pressure to tighten than it would otherwise.
There would be less reason to worry about any of this if the Fed had adopted an explicit nominal-income target, or adopted one now. Under such a target, the Fed would commit to adjusting monetary policy so that nominal income grew at, say, 5 percent a year (roughly the rate at which it expanded in the decades of the “Great Moderation” before the crisis). If the economy grew by 3.5 percent in real terms, for example, inflation would run at 1.5 percent. The Fed would also commit to correcting for past mistakes: If it let nominal income grow by 6 percent in one year, it would subsequently keep its growth below 5 percent in order to keep the long-run path as close to the predicted one as possible.
The Fed would not have had to amass as many assets as it now has if it had followed this policy. If markets had expected the Fed to keep nominal income on a steady path, the demand for the safety of money would not have risen as much as it did, so there would have been less need to increase the supply. As Australia shows, a “looser” monetary policy can lead paradoxically to a smaller money supply. It has kept nominal-income growth relatively steady, and its monetary base is smaller compared with its economy than is that of the U.S. (or of most other countries that let nominal income crash). It has also avoided the last two recessions that hit the rest of the developed world.
Even after the initial decline in nominal income, the adoption of a credible commitment to nominal-income targeting could have kept the Fed from having to buy so many assets by reducing the demand for money balances. (It is partly because of this sort of effect that studies find that expectations of future nominal income are a strong determinant of current nominal income.) And nominal-income targeting could moderate any future money-demand shocks in the event that, for example, Europe collapses.
A credible nominal-income target would also aid the Fed’s unwinding by moderating the decline in the demand for money balances during a recovery. If markets don’t think the Fed will allow nominal income to grow 10 percent a year for the next decade, a bubble psychology is less likely to set in.
And it should be easier for the Fed to commit to a credible nominal-income target than to an inflation target because the former would not require the perverse actions that the latter would. If the target is 5 percent, a nominal-income-targeting central bank will, in a year when the economy grows by only 1 percent in real terms, let inflation rise to 4 percent. When the real economy grows by 4 percent, it will let inflation sink to 1 percent. An inflation-targeting bank would have to adopt tighter money during the bust and looser money during the boom. So it will either have to make the business cycle more extreme or fudge its target.
The worriers are wrong. The Fed can and should unwind its balance sheet without hurting the economy — especially if, at long last, it adopts the right monetary policy at the same time.
– Mr. Ponnuru is a senior editor of National Review. Mr. Beckworth, a former international economist at the Treasury Department, is an assistant professor of economics at Western Kentucky University and the editor of Boom and Bust Banking: The Causes and Cures of the Great Recession.