If you have read, written, or watched papers about monetary policy without money, it may be time to watch this: (Thanks @SEDmeeting for the invitation and @EconomiaDiTella for the hospitality.).

The one and only.

In case you are interested in the details, our latest draft is here:.

This is how the aggregate demands for reserves that we estimate for the United States between 2017 and 2019 look like:

Hi @Isabel_Schnabel this is such a great thread. Thanks for sharing it. By the way, @gmnavarro and I could have used it almost word for word to write our introduction section. (We may reference some bits of your speech in our next revision!).

(4) We use the quantitative theory to: (a) estimate the aggregate demand for reserves in the United States, and (b) develop “navigational tools” to guide the implementation of a fed funds target.

(3) Show that the quantitative theory is flexible enough to match a rich set of calibration targets (mostly on loan *quantities*), and that it matches a wide range of observables on loan *rates*.

(2) Develop a prototypical search-based theory of the fed funds market into a quantitative model with heterogeneous banks (in terms of payoffs, market power, degree of centrality in marketmaking, and frequency and size distribution of payment shocks).

In the paper we do four things: .(1) Document a wide array of micro (bank-level) and marketwide facts about the fed funds market. We think this empirical work could be of independent interest for anyone who wants to learn how the fed funds market operates.

The fed funds market has operated with a very large supply of reserves for over a decade, and now nobody really knows how the aggregate demand for reserves looks for lower levels of reserves. This is Question 2 that @gmnavarro and I want to answer.

But, as it turns out, nobody seems to know exactly how ample is "ample". This led @gmnavarro and I to want to answer the following:.Question 1: In terms of the figure above: what are Q' and Q_{1} in dollars?.

The Fed intends to continue operating a floor system with "ample reserves, in which active management of the supply of reserves is not required".

The Corridor System was used when reserves were scarce, and the fed funds market was operating on the steep range of the aggregate demand for reserves. The Floor System is what you use when reserves are so large that you are on the flat range of the demand.

If you picture a downward sloping demand for reserves (with the FFR in the vertical axis and the total quantity of reserves in the horizontal), and a vertical supply (controlled by the Fed), the two operating systems look like this:

The two most popular operating frameworks are the "Corridor System", used by the Fed before the Great Financial Crisis (GFC) and the "Floor System", which is what the Fed has been operating since the GFC.

In Fed jargon, an "operating framework" is a consistent usage of these instruments. Some operating frameworks tend to rely more on Q-based instruments, others, on P-based. E.g., in the modern US: mostly Q-based prior to 2008, and P-based since then.

The implementation of the FFR target relies on two kinds of instruments: Q-based (e.g., repos, outright asset purchases), or P-based (administered rates, e.g., discount-window rate (DWR), interest on reserves (IOR), or the offering rate on overnight reverse repos (ONRRP)).

In reality, the FOMC understands that the FFR they voted won't "just happen", so they issue instructions to the Open Market Desk at the FRB-NY to create conditions in reserve markets to maintain the FFR within the target range.

In macro we talk about "the (target) policy rate" (e.g., the fed funds rate (FFR)); think about how it should be set, how changing it affects the economy through a myriad of transmission channels, etc. In the models, once the central bank chooses a FFR, the rate just happens.

Flight from LA to NYC. The guy in front of me does not believe in the social norm of not reclining seats, so I cannot watch movies. This may be a good time to write something about his work we are doing with @gmnavarro.