How About Pegging a Long Term Interest Rate ?
As often, Paul Krugman has an extremely interesting thought. Discussing the sudden end of the Swiss Franc Eur peg he wrote
This in turn helps us put the explicit exchange rate target into the right slot: it was about making QE effective through commitment, so that you got the maximum impact on expectations. Actually, the success of the currency program suggests that other central banks might want to try things like setting a ceiling on some long-term interest rate.
I think the paragraph is pretty clear but will try to explain a bit more. There are two huge advantages of a peg. First the rule for maintaining one is very simple — just sell at a price as much as people want to buy and buy as much as they want to sell at that price. Also, whenever markets are open, people can see if it is being maintained . We can see that the Fed is sticking to its declared target federal funds rate immediately. It always does. I’m sure the Federal funds rate is perceived by some (of the those who know what it is) to be an instrument under the administrative control of the Fed.
The Fed could declare that the 10 year constant maturity treasury rate shall be no higher than 1% and then make it so by buying enough notes and bonds.
I see two questions. First which long term interest rate ? Second, what might go wrong ?
The interest rates the Fed would most like to set are interest rates on risky assets such as mortgage bonds or at least low grade corporate bonds. The problem here is that the say the Baa corporate bond rate is an average of the rates on many different bonds. If the Fed were to promist to buy any of them at a yield of 2% say, then it would buy the worst Baa bonds. Pegging one price for a bunch of different things is the perfect way to achieve adverse selection. I think that it is much more practical to set a long term Treasury rate.
Now the 10 year constant maturity rate isn’t the return on a particular note or bond. It is an index calculated with a complicated formula. However, it is calculated with a standard formula and can be recalculated very quickly. All the different Treasury securities the returns on which are involved in the calculation are equally safe.
To simplify, I am now going to consider a silly but simple target in which the Fed puts a ceiling on the return of a particular bond which I will call the target bond. That is equivalent to setting the price of that bond and this can be achieved buy buying or selling unlimited amounts at that price.
OK so what could go wrong ? Here I think the problem is that it might be that to keep the price of the target bond at the desired floor (so the yield at the desired ceiling) the Fed would have to buy exactly 100% of the target bonds that have been issued. Then the market for the target bond would freaze with the Fed offering to sell it for a price higher than anyone is willing to pay.
The effect of the simple policy would be to eliminate an asset. Returns on very similar assets which are not bought and sold by the Fed might be very different.
This silly result is partly due to my considering the silly so simple I could do it policy. But I think that it illustrates a real problem with setting a ceiling on a long term interest rate -the yield curve is flexible and pushing on one part of it doesn’t have to move all of it. I think there is a tradeoff between setting a simple well defined long term interest rate target the monetary authority can actually reliably hit and influencing a large part of the huge set of long term interest rates which matter.
I think there is a lot of devil in the detail of choosing “some long-term interest rate”.
Your danger arises solely from the needless simplicity of targeting just one (or a few) bonds. The Fed could draw a rate curve ceiling – 1% for 5 years, 2% for 10 years, 3% for 20 years, etc – interpolating between the points based on a formula. Any bond issue can easily be placed on the graph based on it’s current duration and price. Any issue above the curve is purchased until all issues are below the curve. If the Fed announces this policy, all issues will immediately jump down to the curve or lower (via expectations) and there will be little need to actually buy.
This might be difficult to work out with a pencil, but it’s trivial with a computer (even Excel) so there is no difficulty in implementation. No there is also no difficulty of creating weird bends in the curve, nor will the Fed need to buy all the bonds at any maturity date.
The rate curve target is easy to explain (it’s one graph) and easy to implement. There may still be dangers, but you haven’t identified them yet.
Why doesn’t the Fed have these problems with short term rates?
I think I know what you’re getting at but consider these two blogposts:
http://www.nextnewdeal.net/rortybomb/did-federal-reserve-do-qe-backwards
“But what if the Fed had done that backwards? What if it had picked a price for long-term securities, and then figured out how much it would have to buy to get there? Then it would have said, “we aim to set the 10-year Treasury rate at 1.5 percent for the rest of the year” instead of “we will buy $45 billion a month of long-term Treasuries.”
This is what the Fed does with short-term interest rates. Taking a random example from 2006, it doesn’t say, “we’ll sell an extra amount in order to raise the interest rate.” Instead, it just declares, “the Board of Governors unanimously approved a 25-basis-point increase in the discount rate to 5-1/2 percent.” It announces the price.
Remember, the Federal Reserve also did QE with mortgage-backed securities, buying $40 billion a month in order to bring down the mortgage rate. But what if it just set the mortgage rate? That’s what Joseph Gagnon of the Peterson Institute (who also helped execute the first QE), argued for in September 2012, when he wrote, “the Fed should promise to hold the prime mortgage rate below 3 percent for at least 12 months. It can do this by unlimited purchases of agency mortgage-backed securities.” (He reiterated that argument to me in 2013.) Set the price, and then commit to unlimited pur”
Policy used to be conducted this way. Providing evidence that there’s been a great loss of knowledge in macroeconomics, JW Mason recently wrote up* this great 1955 article by Alvin Hansen (of secular stagnation fame), in which Hansen takes it for granted that economists believe intervention along the entirety of the rate structure is appropriate action.
He even finds Keynes arguing along these lines in The General Theory: “Perhaps a complex offer by the central bank to buy and sell at stated prices gilt-edged bonds of all maturities, in place of the single bank rate for short-term bills, is the most important practical improvement which can be made in the technique of monetary management.”
*http://slackwire.blogspot.com/2014/10/alvin-hansen-on-monetary-policy.html
I am very disappointed in this post. I clicked on the link to it (h/t Mark Thoma) expecting to find some kinky porn for economists. Not only is it not kinky, it’s not even porn. It does seem to have something to do with economics though. So 1 out of 3.
Kevin you are the devil in the details. Yes your proposal makes sense.
Peter. One might imagine that the Fed targetting the federal funds rate is moving just the shortest end of the yield curve. Some have argued that this doesn’t work, because it is medium and long term rates which matter for aggregate demand. The strange fact is that, when the Fed drives up the federal funds rate to a huge level, long term rates also become huge (not as huge but huge). The rate on standard 30 year mortgages was gigantic in the early 80s reaching 18.45% in October 1981.
I think it is hard to understand why this happened. I think it was due to extremely irrational expectations. Of course the duration of a standard 30 year mortgage isn’t 30 years — the borrower has the right to refinance. I think home buyers must have planned to refinance — otherwise it would have been crazy to borrow long term at 18.45 %.
I struggle with the issue here
http://angrybearblog.strategydemo.com/2014/12/home-price-expectations-and-residential-investment.html
In any case, the data say the yield curve is much less flexible than I think it should be making the problem less difficult than I would tend to guess. In any case, it can be solved it’s just the FOMC minutes would be very long (describing yields as a function of maturity including those damn coupons).
.
When the Fed stopped buying (or tapered), the price of long term Treasuries ROSE. Why would that be?