Sheriff, this is no time to panic.
This is a perfect time to panic!
In A Fed Primer, the mechanics of a Quantitative Easing (QE) balance sheet was shown from a t-account perspective. Although one, and in fact many did, expect QE to be inflationary in proportion to the amount of monetary base that was created, it was not. As was shown in the prior article, it is not the monetary base that drives inflation but what has been labeled "effective money." Effective money and inflation have direct impact on inflation linked securities (TIP) and US Treasuries in general (TLT). In addition, rates and their economic impact drive stock market results (SPY). It is relatively widely expected that the Fed will slow the pace of QE at the next meeting on September 17/18th. It makes sense to review the options that the Fed has and what that means for the Fed's balance sheet and different measures of the money supply.
Run the Ol' Switcheroo
There are essentially three strategies to decrease the Fed balance sheet. Of course, there are combinations and variations, but there are three core strategies: letting the securities mature while financing them, selling securities and "swapping" with the Treasury. The last seems magical in that it at once deals with the Fed balance sheet and retires Treasury debt all without any market impact. Proceeding through each mechanically and following the credits and debits will show that there is no magic bullet.
Our example expands an economy that originally had $100 in assets to one that has $300 in assets via UST debt issuance and QE. Returning to the balance sheet and t-accounts, the monetary state of the model economy looks like this (from the prior article):
Time 2 (Fed purchase):
Let's begin at the end and consider the case for the magic bullet, that is, to swap the Fed's notes for the US Treasury's liabilities. The first thing that is fishy about this becomes evident when trying to figure out exactly what to put down in the t-accounts. The Fed is going to hand over its $100 Note to Treasury but is not going to receive anything back. Treasury will not hand over currency to exchange for those securities because it does not have the cash (we are still running a deficit); an IOU from Treasury is equivalent to the original note and cash raised with larger auction sizes is equivalent to the Fed selling the security to the market. A rational, economically motivated party is not going to hand over an asset without appropriate compensation. Some journal entries are going to have to be created for both Fed and Treasury in order to ensure that balance sheets actually balance (by definition, they have to). If that is done, that is a de facto merger of the Fed and the US Government. And that is a permanent monetization of the debt. Quantitative easing is a monetization of debt, but it is a temporary situation. If the Fed forgives the debt, then QE becomes a permanent monetization of the debt. At that point, one can consider that a US government debt default has occurred. It is not a principal default but a default via inflation. Inflation has been historically the most common type of government default (see Reinhart & Rogoff). That would be a very qualitative game changer.
No Taper Tantrum
The distinction between the temporary and permanent states is absolutely critical. Unlike prior QE operations from the Fed that were limited in scope, the Fed set out to condition the markets by keeping the purchases open ended. For this latest round of QE, the purchase amounts per month were announced, but at the outset the Fed stated that purchases will continue until the economy was deemed not to need it. Hence the moniker QE-Infinity was given to the operation.
A rational bond investor would have a set of probabilistic outcomes that would include a range of scenarios. First is the obvious; that is, the Fed is buying $85B of bonds each month. That is going to force asset prices higher, particularly for USTs and mortgages. Second, it turned out that the Fed became a proportionally larger buyer as greater tax revenues caused the deficit and auction sizes to shrink. Finally, the probability that the Fed would permanently monetize US government debt simply had to increase with QE Infinity. There had always (and still remains) a danger that the Fed would simply monetize the debt; the indefinite, long term nature of the operation increased that chance. The probability for monetization made inflation protected securities most attractive. Real yields on TIPS collapsed, with 10-year real yields reaching as low as -0.9%. This matches the market action: breakeven inflation (10 year note less 10-year TIP real yield) moved from 2.3% to 2.6% when QE3 was announced; it got as tight as 2.0% after tapering was announced.
When the FOMC in June announced QE was going to be wound down and eventually ended (contingent upon economic data), two out of the three buying incentives ended. It was clear that the probability for a permanent monetization had dropped dramatically. This is why this round of QE is different than the others. It had been designed as an ongoing, open ended operation. The others had not and in the earlier rounds, the bond market had already been looking past the operation even before Fed purchases had begun so that yields did not drop as anticipated. The market couldn't do that in this case; it had to address QE for the long term right up until Bernanke said "whoa." Whatever schedule of tapering now occurs is now a quantitative question and of far less importance than the qualitative announcement that QE was slowing and then ending.
The end result is that it makes perfect sense for yields to back up to the range where they were prior to any QE. The range for real yields on TIPS ($TIP) was throughout 2010-2011 in an extended 60-130bps range. For 10-year notes, the range was a wide 2.5-3.5%. For 30 year bonds ($TLT), the range was 3.5-4.5%. In all cases, current rates are almost perfectly in the middle at 85bs, 2.93% and 3.87%. No panic here.
While the Fed cannot realistically forgive Treasury's debts, it does have four operational strategies for unwinding:
- First, it can simply allow securities to mature. Then it needs to eliminate the cash just as it created that cash previously for QE. This works so long as the money from the maturing securities roughly matches the money contraction needs of the Fed.
- Second, it can act in the repo market. It exchanges the long term notes in its portfolio for cash from banks. Repo tends to be short term, mostly overnight, transactions. Although repo is a buy-sell arrangement, it is a de-facto financing agreement. The Fed will be borrowing money from banks in a collateralized loan and will be paying the market rate.
- Third, it can attract excess reserves by paying a rate of interest higher than that prevailing in the Fed Funds market. At some point it may be necessary to compete with investments that may become more compelling than the Federal Funds market.
- Fourth, it can sell the securities in its portfolio. As in the first case, it then destroys the cash it receives to offset the cash that it created before. The Fed will either record a gain or loss on those securities.
The first three choices are all centered on allowing securities to mature and adjusting any cash balances in the interim via repo (2) or excess reserves (3). Should excess reserves be removed from the Fed, it is possible that the Fed can offset those losses via "printing" additional reserve balances, i.e., expanding the monetary base further.
It is worth mentioning that QE can be viewed as the Fed taking High Quality Collateral (HQC) from the market (USTs and Agencies) and converting it into cash. The Fed, purposefully or not, has sterilized its QE operations. That is to say that the cash has been parked at the Fed in the form of excess reserves because the Fed has paid interest on those balances higher than other equivalent alternatives (think repo and Fed funds). That money has not found its way back into the economy (see Fed Primer). Further, by taking high quality collateral from the market, it has likewise shrunk the pool of assets that can be used to back multiple levels of loans (re-hypothecation).
The Fed desire is that unwinding should not negatively impact economic growth nor should it accelerate inflation. It more or less follows that the Fed wants to keep the supply of effective money stable. That does not mean that with a stable effective money supply that inflation will be zero; it is reasonable to expect it to be within a controllable range. There are three main dangers to such a benign outcome: additional HQC entering the market, the Fed's conflict of interest and the potential for inflation beyond the Fed's control.
The Fed has recently announced a Fixed Rate Reverse-Repo Facility (FRRRF). The announcement implies that the Fed is looking toward repo financing (#2 above) rather than excess reserve financing as it is doing now. Part of the reason for this is the spread between the overnight Repo market rate and the Fed's interest rate target. A good discussion can be found on Scott Skyrm's blog and FT Alphaville. There is a widely held view that there is a collateral shortage and this facility is expected to help alleviate that shortage as well as bring multiple overnight lending rates into line. Overall the market reception has been warm to this announcement. As discussed above, though, is that QE converted HQC to cash and that cash has found its way back to the Fed until now. Although it is impossible to determine what the result might be, it stands to reason that if the Fed expands the availability of HQC, there is the potential for effective money to expand: the collateral that gets used in the FRRRF can be re-hypothecated although the usage rate is unknown. There is reason to believe that re-hypothecation will not be as efficient as prior to Lehman's bankruptcy because some hedge funds will not allow re-hypothecation in prime brokerage accounts due to the difficulty in obtaining funds if a counterparty in the chain fails. Similarly, Europe may limit collateral reuse. It still seems likely that at least a gradual but steady increase in effective money will occur if increasing amounts of HQC enter the market.
The Fed is running a $3.5T (source: Fed) portfolio which is still growing and will continue to grow even with the taper. Based on Table 2 in H.4.1, the bulk of the Fed's UST holdings have a maturity of 5-10 years, with a roughly equal amount maturing earlier or later. Nearly its entire mortgage portfolio matures after 10 years. The Fed is lending long and borrowing short. It is doing so with much greater leverage than the banks that nearly collapsed the financial system five years ago. The Fed is capitalized with about $55B (source Table 1, FRB 2012 annual report). That makes the Fed leveraged to the tune of about 63X ($3.5T/$55B). That level of leverage means that a 1.6% portfolio loss (1/63) wipes out the Fed's capital. Bloomberg's BUSG index corresponds to a roughly 7 year maturity (5.7 year effective duration) Treasury/Agency portfolio and is a nice comparable to the Fed's balance sheet; it is down 2.5% since late May already. Even so, the Fed will not go bankrupt; any shortfalls are covered by the Treasury in the same way but opposite direction as when the Fed transferred gains to the Treasury. Nevertheless, there will be strong incentives within the Fed and at Treasury to avoid realizing losses, particularly large ones. In addition to pronouncements from the Fed suggesting that it will not be selling its portfolio outright, the losses already incurred by the portfolio stack the probabilities toward the Fed holding until maturity.
Therein lays the conflict. Unlike other banks, the Fed is able to control (not completely) short-term rates. On the one hand, the Fed has the responsibility to raise rates if and when the economy improves. On the other hand, the Fed is running a portfolio whose value and cost of financing is directly linked to the path of short term interest rates. Raising rates will directly cost the Fed multiples of its interest expense. Each 25bps costs the Fed approximately $5B per year (I used $3.6Bx25bp which is $5.4B; the Fed notes total interest expense of $1.234B/quarter or $4.9B with rates currently ~25bps). Even with the most responsible leadership, the undeniable fact that acting to raise rates will be in the mind of those making the decision. It is unavoidable. Behavioral economics teaches that people will find rationales for acting in their selfish interest and the selfish interest of the FOMC is not to print large losses. The FOMC is subject to the same human characteristics and will likely find and believe theories which support keeping rates lower than would otherwise be prudent. FOMC members are every bit as subject to conflicts of interest as any other person. That means that the portfolio and its profit and loss statement indicate greater than normal risks of inflation if the economy continues to improve.
If the economy continues to improve but is still dogged by lack of collateral, it is entirely possible that the marketplace will become comfortable with alternative forms of collateral. This does not immediately seem likely, but the market has adapted in the past. Wall Street has great incentive to find ways of creating debt. Treasuries have been around for a long time. It has only been recently that re-hypothecation or the shadow banking system became of a size comparable to M2. As noted in the May 2013 TBAC Discussion, "collateral moneyness is pro-cyclical." That means that if the banking system wants to make loans and needs to source money, it will find a way to do so and find the economic justification for it later. If the economy heats up and traditional sources of money are not available, then the banking system will find some way to expand the money supply. This is another way of stating that the definition of what is "effective money" will change over time. This means that a quantitative definition of effective money will also need to evolve. This is a sort of "variable drift" and is one of the difficulties of strictly quantitative modeling. The market may find another source of debt creation that has not yet been anticipated. We know that the market will find a way to supply leverage and to create money when there is a demand for it. Traditionally, the Fed and other regulators have been slow to acknowledge shifts and slower still to address them.
Conventional wisdom in the market right now views a tapering of purchases as a tightening of credit. That wisdom should be respected but it is important to ask the correct questions. The first question is what impact a tapering will have on the supply of money. The second is how much impact a rise in long-term rates have on the real economy.
Tapering may not have the anticipated effect on the economy that many think. The Fed has created money in order to purchase long-term securities. Because the Fed has been paying balances on excess reserves, that money has not found its way into the real economy. It is entirely possible that as the Fed unwinds its portfolio, the additional collateral that will become available to the market will expand the supply of effective money. It may not make an immediate impact but seems likely to cause a gradual but steady increase in effective money. That would be stimulative and, potentially, inflationary.
On the other hand, higher rates clearly are a tightening of financial conditions. The absence of the Fed as a buyer of longer term debt represents less demand. Nevertheless, rates represent an expectation of the path of short-term rates over time. It seems clear that the Fed will be hesitant to raise rates as the economy improves. First, the Fed does not want to be the cause of a reversal in the economy. Second, due to its large portfolio, it has a conflict of interest and its institutional bias will be to keep its funding source cheap. Interest rates, even out to ten years, will be contained for both of those reasons.
The answers to both questions point toward inflation ahead over the longer term. Prefer TIPS over regular US Treasuries. Not only does this make sense based on Fed operations but the spread is at a reasonably attractive entry point based on the histogram of historic spreads. In fact, the histogram would look significantly more attractive without late 2008/early 2009. If one believes that the US is on the path to normalcy, then this trade should work well.
Additional disclosure: I am long TIPS via WIW. I may change my position or my view based on additional information or analysis.