1993-redux?

Since the introduction of QE, the level of excess reserves has had a reasonably tight correlation with the level of short-term interest rates (reassuring for the monetarists!). There are reasonable grounds to expect these to rise somewhat, at the margin. Firstly, the recent discount rate hike has removed some of the safety net for the banking system and is likely to be hiked further as part of the normalisation process. Under this assumption, it is reasonable to expect that the current ~$8bn of discount window borrowing will fall as private financing elsewhere starts to appear more attractive.

Secondly, the Treasury recently announced that the Supplementary Financing Programme (SFP) will be reinstated to $200bn. By way of background, this programme was introduced in Autumn 2008 to help the NY Fed regain control of the Effective Fed Funds Rate at a time when the Federal Reserve was undertaking exceptionally large interventions in the financial market (it effectively took the entire interbank funding market onto its balance sheet). In Autumn 2009, the Treasury was running very close to the debt limit at a time when the Administration was trying to steer Healthcare reform through Congress. To avoid introducing further potential Congressional grandstanding, they decided against asking for a formal increase in the debt limit until the Healthcare measure had progressed. Instead, they announced that they would run down the SPF to provide bridge finance which, in effect, put an additional $200bn of excess reserves into the banking system, contributing to the last-gap jump in the Liquidity Trade in November. The re-announcement of the programme represents the reversal of this, as Congress approved an increase in the debt limit back in January.

Now, while neither of these moves represent an explicit tightening of monetary policy under the FOMC's current framework of targeting the Federal Funds Rate (and several FOMC members immediately came out insisting that both moves were not tightening), a $200bn+ reserve drain will inevitably have some form of impact on market rates. The below chart shows Excess Reserves in the banking system (inverse scale - Brown line), the Effective Fed Funds Rate (Green line) and the 12m T-Bill rate (White Line) and, as per above, the brown line is about to turn higher. Many market participants viewed the recent Fed actions as representing a tightening, and many insisted it was not. Interestingly, those that viewed it as a tightening were generally not core-participants in the Rates market, but rather were concentrated in other asset classes. Rates market participants immediately faded the immediate sell-off in the front-end on these moves. So who is "right"? Is this a case of "watch what we do, not what we say"?.

One of the most remarkable observations, with respect to interest rate markets, has been the stability of front-end carry trades following the view that policy rates will be on hold for a very long time. In fact, the best risk-adjusted trade of 2009 was long EDZ9 and, so far, EDZ0 is making a strong claim to the 2010 title itself. This view has been supported by continued dovish rhetoric from central bankers and academics (most notable the Rogoff & Reinhart study that highlights a significant amount of evidence that implies that post-financial crisis recoveries tend to be L-shaped) and suggestions that global output gaps are very wide. All of these points are valid to some extent, but the price action, as well as the outright level of the front-end, is beginning to overprice these issues and it is becoming obvious that the trade is no longer fundamentally-driven and is now a pure momentum-driven trade. I will attempt to show this below.

The below chart shows a forward-looking Taylor Rule estimate using the average economist estimates for core PCE inflation and unemployment. These estimates are consistent with an L-shaped recovery and very gradual fall in the unemployment rate where the size of the output gap exerts downward pressure on core inflation. Under these conditions, the forward Taylor Rule suggests that the policy rate 1yr forward should be -1%. This, of course, does not take into account the effect of QE upon monetary policy. Calculating the impact of QE is obviously exceptionally difficult and, perhaps, somewhat subjective, but several months ago Goldman Sachs did try to do this. By utilising the equivalent easing of their Financial Conditions Index of a 100bps Fed Funds cut, they were able to estimate the quantity of asset purchases that would be consistent with such a move by comparing the QE-effect upon their Financial Conditions Index, coming to the conclusion that this quantity was about $1trn. By the end of Q1, the Fed will have purchased a cumulative $1.75trn, which under the above assumption, can be thought of as roughly equivalent to 175bps of additional easing.

Adding this into the below chart, we can see that the optimal Fed Funds rate 1yr forward should be 0.75% if the assets held on the Fed's balance sheet remain unchanged. This is pretty much where Fed Funds futures are priced, however that does not take into account term premium (which according to back Eurodollars is about 15bps/year), so the equilibrium forward rate should be something like 0.9%. This is somewhat back-of-the-envelope analysis, but it is illustrative that expecting further extension in the front-end is purely dependent upon the growth outlook deteriorating further. Also interesting is the average economist expectation for the path of the Fed Funds rate, which is for it to sit at 1.25% in one year's time. The point here is that the rates market already prices in an L-shaped recovery.

The next thing I want to draw attention to is speculative positioning (see chart below), which now sits at its longest since the CFTC began collating the data in 1995. As market-wide liquidity is lower post-crisis, on a liquidity-adjusted basis, the net-length is effectively larger than it appears. The performance of front-end carry trades has attracted further momentum buying - in particular, since the beginning of March, Open Interest in the front seven contracts has increased by over 160k contracts, or about ~$4m/bp in PV01 terms, with brokers reporting that the price data shows that the majority of this volume has been traded on the offer side of the price, suggesting that the market is increasing its length further.

But it is not only speculators that are playing the carry trade. The steepness of the curve has encouraged Commercial banks to pile into Treasuries, now holding the largest amount on record (see first chart below), with an average duration of about 5.5yrs. Relative to Nominal GDP (second chart below) the holdings are not quite as extreme as they were in 1993, but given bank balance sheets are in much worse state as a result of the current crisis, the effect of losses from a Treasury market sell-off will be far more dramatic. It is also worth noting that the peaks in this ratio in both July 2003 and March 2004 at levels very similar to those at present were respective peaks in the rate market and met with very aggressive sell-offs.

The remaining rates market participants are Bank Swap Desks for whom there are no positioning data, but anecdotally I am led to believe that they also have the carry trade on in significant size. It is worth noting that the swap market is significantly larger than it was in the early-1990s (in 1993, according to ISDA, outstanding Interest Rate Derivatives totalled $8.474trn, while in 2009 they totalled $414trn.). Indeed, from a balance sheet perspective, it is more efficient to express this trade in swaps rather than bonds, so this would suggest that the true size of the carry trade is much larger and that the UST holdings data understates the true size of the carry trade.

So the point I'm trying to get across is that the front-end is over-shooting as a function of momentum (perhaps CTAs are in on the action) and speculators chasing carry-returns. It's an accident waiting to happen.