by Benn Steil -- co-authored with Emma Smith

How the Fed Should Tighten Monetary Policy

On January 27, the U.S. Federal Reserve held interest rates steady and, in a modest nod to a market that has been consistently more fearful about the economy than the Fed itself, dropped a line from its December statement saying that the risks to the outlook were "balanced."

Clear signs of a deteriorating global economy reinforce our view that Richard Koo, chief economist at the Nomura Research Institute, is off the mark in his critique of the Fed's approach to monetary policy.

In his recent article, Koo argued that the Fed is tightening monetary policy with the wrong tool. Rather than hiking interest rates (by raising its target for the short-term Fed funds rate), he writes, the Fed should sell bonds from its balance sheet, which has been inflated by years of quantitative easing. Indeed, at a size of $4.5 trillion, the balance sheet is now five times larger than it was before the economic crisis. Selling off assets would shrink banks' excess reserves, he continues, reducing their incentive to lend. Rate hikes have the same effect, but Koo thinks that asset sales are better: if bond yields rise too much, or the economy turns sour, the Fed can just announce an extension of the zero-rate policy to "calm the markets."

There are at least two major problems with Koo's argument. The first is that he offers no numbers, and therefore no guidance as to how much tightening balance-sheet reduction would achieve relative to rate hikes. The second is that his logic is backwards.

Let's first look at the numbers.

Longer-term bond yields in the market, which the Fed is looking to nudge upward, can be separated into two components. The first is the expected path of short-term interest rates. Changes in the Fed's policy rate, its guidance about the pace of rate hikes, and the economic outlook all affect the expected path of short-term interest rates.

The second is a term premium, which is the compensation that investors require for holding a longer-term bond (as opposed to rolling over short-term bonds for the equivalent length of time). The term premium fluctuates with supply and demand for particular bonds. The Fed's accumulation of longer-term U.S. Treasuries has been a big source of demand and has therefore lowered the term premium on longer-term Treasuries. (Former Federal Reserve Chairman Ben Bernanke explains this well on his blog.) Fed economists estimate that yields on ten-year Treasury bonds would be around 70 basis points higher than their current level of around 2 percent had the Fed not bought such bonds to begin with.

In short, when the Fed buys and sells bonds, it affects the interest rates on such bonds. So does the Fed pushing the Fed funds rate up or down. As a rough guide, a rate hike of 100 basis points is historically associated with a 25-point rise in ten-year Treasury yields.

The Fed expects to raise the Fed funds rate about 100 basis points, from the current rate of 0.375 percent, over the course of the coming year. This should, then, be expected to push up ten-year Treasury yields by about 25 basis points. The numbers above suggest that the Fed could achieve the same rise in ten-year Treasury yields by letting its maturing bonds roll off and selling an additional $600 billion worth.

Should the Fed sell bonds, then, instead of raising rates, as Koo argues?

No. If the Fed sells bonds rather than raises rates, and the economic outlook sours, as it has since December, Koo would have the Fed rely on the weak gruel of forward guidance to stimulate spending and investment. That is, the Fed will just tell the markets that it intends to keep rates at zero for a long time.

Experience with forward guidance in the United Kingdom and Switzerland, however, has been miserable. Words don't matter nearly as much to markets as action. So if the Fed should find the economy headed for trouble, it would much rather be in a position to cut rates than to have to jawbone the markets - or, worse, launch a politically toxic fourth round of quantitative easing and buy back the assets it sold.

In short, Koo has this wrong. Although it would have been better not to have raised interest rates in December, the Fed clearly made a less bad choice in raising rates than it would have by selling assets. The time for asset sales is when the economy is booming - that's when having additional tools for tightening policy becomes valuable.