I was reading a very informative post on Slack Wire recently regarding QE and why it’s difficult to move long-term rates. The basic idea is that the price of an asset and its yield follow opposite paths. A fall in interest rates (and as a result of yields) will lead to higher prices and capital gains for asset holders. But in an environment of low nominal interest rates (based on the Fed commitments interest rates will remain very close to the zero bound for an extended period of time), an increase in rates can easily create very large capital losses for players who finance their asset holdings by borrowing funds (especially by repoing the assets themselves). As a result, the Fed has to permanently move the “expected” Federal Funds Rate path in order to be successful in lowering long-term interest rates. Only committing to a future short-term rates path is not enough, but actual purchases (which move capital gains to the present) are needed in order to “convince” the market that the commitment is credible, by providing it with a nice profit and a zero maturity interest bearing account.
In such a context, one can extend the above by taking into account Interest on Reserves. Since the Fed will use IOR as the basic tool to steer interest rates in the future (since excess reserves will probably remain much higher than needed to achieve an interest rate target by manipulating reserve quantities), it is basically moving any future capital loss risk on its own balance sheet. The higher the QE holdings (compared with long-term Treasuries held by banks and dealers), the higher the “insurance” that any interest rate hikes will not hurt the financial sector since IOR income will more than compensate for any unexpected rate increase. That allows the Fed (and the financial sector) to bring lower long-term rates forward and influence private sector borrowing decisions (to the extent they are affected by interest rates and not income expectations).