What many of us commonly refer to as QE is known around the Fed as large scale asset purchases, which are actually just one part of what the Fed calls "credit easing." Under current policy, the Fed buys $40 billion of agency paper and $45 billion of Treasurys.
The effectiveness of these asset purchases is still controversial. Because they have been accompanied by other forms of easing—such as commitments to keep rates low until economic goals are met—it is difficult to isolate the effects of the purchases themselves.
What's more, even if we think we know what the asset purchases accomplish, we might not know how it does the job. Is it the lowering of supply of safe financial assets that pushes yields down? Or are the purchases really just another form of communications strategy, in which the effect on asset supply is less important than the effect on market expectations?
A 2011 paper by a duo at the Kellogg School of Management concluded that asset purchases had caused "a large and significant drop in nominal interest rates on long-term safe assets (Treasury, agency bonds, and highly-rated corporate bonds)." It argued that the effect was indeed rooted in the reduction in the supply of long-term safe assets. The paper also found a significant affect on mortgage-backed securities—but only when the Fed was buying the securities.
For the purposes of our intellectual experiment, let's say the Fed shares this view of its asset purchases. Could there be a situation in which the Fed thinks short-term interest rates should rise, but rates on longer-term safe assets and mortgage-backed securities should continue to be held down? That is, could the asset purchases continue even after the Fed hikes its target rate?
The easiest affirmative answer to these questions begins in the housing market. If the housing market seriously slumped once again, or the Fed believed it would slump as it hiked interest rates, the Fed might well decide that it wanted to keep mortgage rates lower than broader rates.
Let's say, for example, that some time in early 2015 the Fed reaches its targets with respect to unemployment and inflation. In order to prevent unwanted inflation across the broader economy, the Fed would presumably begin to raise rates. If housing were still seen as a particularly fragile sector of the economy, the Fed might continue to purchase mortgage-backed securities in order to prevent the tightening it wants in the rest of the economy from triggering a new mortgage crash.
Turning now to long-dated Treasury bonds, we can also tease out a scenario in which the Fed keeps its asset purchases going while raising rates. One of the dangers with any decision to raise rates will be that the market will misinterpret the announcement as a sign that rates will rise much more rapidly than the Fed thinks its healthy. The yield curve would become far steeper as longer-term rates respond to the expectation that future short-term rates will rise.
(Read More: Warren Buffett Says Bonds are a Bad Investment)
What's more, investors fearing rising rates might engage in an undesirable "fire sale" like selloff of government securities. Since long-term rates have a far bigger affect on the investing and borrowing decisions than the short-term rates directly influenced by the Fed, central bankers could find themselves "losing control" of the long end of the curve.
The Fed could try to avoid this with language in its official statements that would attempt to reassure investors that future rate hikes would be gradual and contingent on strengthening economic performance. This would heavily depend on the Fed's credibility, however, which might be in doubt. Ben Bernanke, after all, will likely have recently retired by the time rates rise. The new Fed chairman may not enjoy the full confidence of the markets.
(Read More: Who Should Replace Ben Bernanke?)
Buying long-dated Treasury bonds could be seen as "backing up" a Fed promise to raise rates slowly. One reason this might work would be that it would give the Fed "skin in the game"—the Fed would take capital losses on its own balance sheet if rates were to rise rapidly.
In 1961, the incoming Kennedy administration launched what's widely known as "Operation Twist." The program was intended to raise short-term rates while lowering, or leaving unchanged, long-term rates. The goal of raising short-term rates was to promote capital inflows and support the dollar.
Most analysts have concluded that Operation Twist was a failure—although one paper co-authored by Ben Bernanke concluded that Operation Twist "does not seem to provide strong evidence in either direction as to the possible effects" of a similar policy. Bernanke and Co. go on to suggest that it might have been more effective if it were a larger operation and carried out over a longer period of time.
None of this is to say you should bet the farm on the persistence of large-scale asset purchases after rates begin to rise. We should all at least acknowledge, however, that we're in uncharted territory here and some of our assumptions about what the "exit policy" will look like may turn out to be mistaken.