Sunday, October 8, 2017

From a Floor System to A Corridor System

So I was looking at the Fed's 2016 annual report and was able to construct the following chart:

Several observations from this figure. First, the sharp growth in the Fed's income is unsurprising as it is a natural consequence of the Fed's QE programs. These large scale asset purchase programs expanded the Fed's assets from around $900 billion in late 2008 to $4.5 trillion today. Moreover, the Fed's portfolio has changed from being mostly short-term treasury securities to one of long-term treasury and agency securities. In addition, the Fed also started paying interest on excess reserves (IOER) to banks during this time. Together, these two developments have effectively turned the Fed into the largest fixed-income hedge fund in the world. Hence, the surge in the Fed's income. 

Second, the Fed's net expenses have also grown rapidly since 2008. Fed officials will point to new financial regulations they have to enforce as the main culprit, but there is arguably a political economy story to the surge as well. Interestingly, this surge accelerated between 2015 and 2016 as seen in the figure. There is no mystery behind this uptick. It is almost entirely the result of the larger IOER payments going to banks as interest rates have gone up. This can be seen in the following numbers:

The IOER payments are expected to grow as the Fed continues to tighten policy. Assuming this tightening does not get ahead of the recovery and cause a recession, future IOER payments should grow as seen in this chart fromThe Economist:

The Fed estimates these payments could hit $50 billion by 2019. And, as a reminder, most of these payments will be going to large domestic banks and foreign banks as they hold most of the excess reserves. This can be seen in the figure below:

So expect the Fed's net expenses to keep growing at a healthy pace for the next few years as it pays out larger and larger IOER payments to large domestic and foreign banks. As I have argued before, this is bad optics and will make the Fed's large balance sheet an increasingly toxic issue for Fed officials. Something will have to give.

Here is my suggestion for the Fed: move to a corridor system. It would still give the 'interest rate control' feature the Fed desires but with a much smaller balance sheet. In a corridor system, the IOER would become the floor for the federal funds rate and the discount rate (or the TAF) would set the ceiling. The figure below shows the difference between a corridor system and floor system. As noted by Stephen Williamson and George Selgin, a corridor system has been working well in Canada with a relatively small balance sheet. 1

Not only is the Fed's large balance sheet bad optics, it really does not pack much punch in normal economic circumstances. So a corridor system makes a lot of sense: it maintains interest rate control, facilitates better politics, and creates a concrete end goal for the Fed's normalization plans that is not disruptive.  Moving to a corridor system should be one of the key objectives for the next Fed chair and the FOMC. 2

1Technically, the Fed has what Stephen Williamson calls a 'sub-floor' system or what George Selgin calls a 'leaky floor' system. Thus, there is a floor beneath the floor, the RRP rate which catches the  financial 'leaks' which make it through the IOER rate. Still, the IOER guides all the other rates and is the main instrument of policy.
2It goes without saying that another key objective for the next Fed chair and the FOMC should be a move towards a NGDPLT. Both this and the move to a corridor system would be complementary. 


  1. Given the recently announced "disinvestment" policy, corridor policy is not a possibility for maybe 5 to 7 years. This depends on the future demand for US currency, in part. The Fed could sell assets, which current FOMC seems strongly opposed to. There's always the possibility that another recession comes along, short rates go to ZLB, and a new round of QE starts, which would further postpone any serious decision about corridors vs. floors.

  2. "There's always the possibility that another recession comes along, short rates go to ZLB, and a new round of QE starts, which would further postpone any serious decision about corridors vs. floors."

    Indeed. And if the Fed sticks to its planned schedule of IOER rate increases (which it erroneously identifies with "normalization" simply because it treats a 1% real ffr as "normal" and then tacks on two points of inflation), the "possibility" starts looking more like a (high) probability.

  3. I don't think the interest payment is a big issue, if the FED don't create that much reserves, Treasury will yield more and that would be the same to the sovereign balance sheet.

    FED is not heading to a standard corridor system or a floor system, more like a binary corridor system, which includes one for cash (reserves) and one for shadow (collaterals).

    Reserves are not the only settlement mean

  4. Suppose that, for one reason or another, the banks holding (excess) reserves with the Fed no longer felt it was the best use of that capital, and began to flood the markets with new loans/other investment. This could happen through a policy change or other means. Wouldn't this huge surge cause massive inflation? I have to think there is serious risk here.