For several months now the US Federal Reserve (Fed) has been describing in incrementally greater detail their plan for normalizing policy when the time is right. Yesterday, Chair Yellen updated the Fed’s sequencing strategy from the June 2011 iteration. The Fed intends to do the following:
For quite some time we’ve expected that the Fed’s sequencing strategy would shift in this direction as support for the MBS/housing market will still be necessary during the early stages of normalization. To the surprise of many, the sequencing (while heavily debated within the Fed) was the easiest implementation aspect of the normalization process. The Chair turned her attention at the press conference to the mechanics of the exit strategy which adds an additional wrinkle to monetary policy. In March I described the two primary tools the Fed had at its disposal to maintain a large balance sheet and still influence short rates (Interest on Reserves “IOR” and the Reverse Repo Facility “RRF”) through the appropriate reserve management. Chair Yellen downplayed the role of the RRF, and a subsequent press release from the NY Fed stated the size of the program would be capped at $300B and the counterparty limit would be capped at $30B. I view this approach as a mistake by the Committee since IOR has not proven to be a sufficient price setting mechanism in the market. The RRF provides a necessary (and thus far, relatively successful) floor on short rates provided the “full-allotment” auction framework is utilized. By imposing a cap on the programs usage (and the counterparty cap), the demand (in the form of reserves in the system) for high quality assets will exceed supply, resulting in market repo (repo conducted away from the Fed) and T-Bills rates trading below the Fed’s intended floor. The cap will prevent money market funds from satisfying their aggregate demand for high quality collateral and require them to transact with the broker/dealer “B/D” community at a lower return. Consider the implications: a money market fund will earn less interest lending cash to an institution with credit risk versus lending to the US Central Bank. The persistence and magnitude of this anomaly will eventually undermine the Fed’s credibility.
Policy makers appear to have gone down this road for two primary reasons:
1) Politics within the Fed – several hawkish regional Fed presidents would like to see a faster normalization (reduction) in the Fed’s balance sheet. They view the RRF as a clear impediment as it will allow the Fed to maintain a large balance sheet indefinitely.
2) Market intrusion – the Fed appears uncomfortable taking on non-traditional counterparties (money market funds), disintermediating B/Ds and crowding out other borrowers during periods of stress. All of these concerns have varying degrees of validity, but they are principally a result of regulations (see below) and can be addressed by program oversight and logistics.
Current RRF usage
In the March blog, I estimated that the RRF would need to have a size of ~$1Trillion to successfully influence rates, particularly at the onset of the normalization process and in a more stringent regulatory regime. To date, the program has averaged approximately $120B with several sharp increases around month- and quarter-end. So, why is the program so small and where is the so called high quality demand? First, the Fed isn’t pursuing an increase in short term rates right now and therefore doesn’t need to remove reserves. Secondly, the demand is coming, as recently enacted regulations, particularly the supplemental leverage ratio “SLR” goes into effect next year and undermines market efficiencies.
Let’s focus on the regulation and its impact to the marginal price setting mechanism in short term markets. The SLR imposes capital requirements on all bank assets regardless of the risk associated with those assets. This rule disproportionately impacts banks with large B/D subsidiaries and discourages high volume, low margin businesses that typically underpin market functionality (match book repo). When the SLR goes into effect, the willingness of B/Ds to provide repo collateral to money market funds will diminish substantially. We already saw evidence of these unintended consequences in June, when the RRF surged to over $330B as B/Ds attempted to window dress their balance sheet over quarter-end to demonstrate the SLR impact. As Michael Kolster, our US Financial Sector analyst highlighted from Goldman Sachs’ 2nd quarter results, “Total assets fell by $56B ($25B from matched repo book) to $860B driven by an initiative to reduce activities with lower returns, including certain secured financing activities.” In 2015 the SLR capital charges will be calculated on average daily balances rather than quarter-end closes. This should be the catalyst for a more pronounced reduction in collateral via the repo market. Otherwise, the marginal capital cost to the B/Ds associated with repo collateral will be reflected in lower earned interest by money market funds which will create an equilibrium level for short rates well below the Fed’s intended target/floor.
The rationale goes, “the Fed will simply adjust their exit strategy if the realized outcomes are not what they intended.” Indeed they certainly have sufficient tools to change direction mid-stream, but “Trial and Error” is not a credible strategy when conducting monetary policy. The Fed’s Credibility has benefited everyone through lower inflation, lower risk premiums and lower yields. I anticipate a refinement of the exit strategy mechanics and an elevation of the RRFs role and more importantly size, in the months ahead. Only then should the market become confident that the path, policy and mechanics of a very challenging exit strategy can be met with a credibleresponse.
 “Breaking up is Hard to Do” http://blog.jpmorganinstitutional.com/2014/03/breaking-up-is-hard-to-do-the-feds-exit-strategy-2/
 Source: NY Federal Reserve, January 1st through September 15th
 Source: NY Federal Reserve, Reverse Repo activity based on a June 30th settlement date