A funny thing happened as every central bank around the world rushed to stimulate their economy by devaluing their currency in a global FX war that is now 7 years old and getting more violent by the day: with bond yields plunging, and over $10 trillion in global debt now having a negative yield, every fixed income investor starved for yield was pushed into the long end of the bond curve where whatever yield is left in the world of "safe" bonds is to be found. As long as interest rates never go up, this strategy is relatively safe. However, a major risk emerges when central banks start tightening.
To be sure, banks have been eager to front-run any concerns about the Fed's rate hike by cheering higher rates as precisely what they need to be more profitable, and the market has so far believed and rewarded bank stocks the higher rate hike odds rose. Just this Thursday, speaking at an investor conference James Dimon said that if short-term and long-term rates were to move up by 1 percentage point simultaneously, 70% of the benefit would come from the move in short-term rates. The reason for this is that even if long-term rates remain under pressure, and the curve flattens further, an increase in short-term rates provides an immediate boost to bank profits. That is because many loans are automatically priced against short-term benchmarks like LIBOR and Prime.
What Dimon did not discuss is the P&L impact from the higher yields and dropping bond prices in the long end of the yield curve. And it is here, in the unprecedented duration exposure that central banks have forced everyone into, that the true risk resides.
How big is the risk? According to an analysis by Goldman's Charles Himmelberg, if rates rise by the Jamie Dimon-referenced 1 percentage point, the market value loss would be between $1 and $2.4 trillion! Putting this loss in context, even the smaller $1trn loss would be over 50% larger than the market value lost in the 1994 bond market selloff in inflation-adjusted terms, and larger than the cumulative credit losses experienced to date in the non-agency residential mortgage backed securities market. And this is only only as a result of a 1% interest rate increase: assuming full normalization of rates to their historical level of 3.5%, and the level of mark-to-market losses climbs to a staggering $3 trillion.
The culprit? The Fed, the same Fed which does not to grasp that by "renormalizing" into the biggest bond bubble in history is assuring massive losses for the financial sector.
The problem is simple: having inflated a gargantuan bond bubble, letting the air out would by definition lead to dramatic consequences not just for bonds but for all other asset classes.
As Goldman shows in the chart below, the growth in total debt outstanding, in constant 2015 dollars, has been unprecedented. The total face value of all US bonds, including Treasuries, Federal agency debt, mortgages, corporates, municipals and ABS, is $40 trillion (Securities Industry and Financial Markets Association). The Barclays US aggregate is a smaller number, $17 trillion, as the index excludes some categories of debt, such as money markets, with low duration. To end up with a more palatable number, Goldman uses the Barclays measure of debt outstanding, although it admits this may lead to an understatement of the total loss potential. Using either measure, total debt outstanding has grown by over 60% in real Dollars since 2000.
It is not just the notional amount of debt that has been relentlessly rising: as Exhibit 3 shows, the aggregate interest rate duration across the bond market has also increased over the past several years, up over 20% vs. the 1995-2005 average level. Longer durations are largely driven by lengthening maturities on the bonds outstanding, as issuers have elected to term out their debt structures. Exhibit 4 shows that the average maturity of corporate bonds issued in 2015 and 2016 is over 16 years, vs. an average of 8.6 years during 1995-2005. The US Treasury has also chosen to lengthen its debt maturity structure, with more use of long duration bonds.
The secular decline in nominal interest rates has also contributed to the drift upward in bond duration.
In 1994, the average yield on the bond index was 5.6%, vs. 2.2% currently. Lower bond coupons means that proportionately more of the bond cashflows now comes from principal, which tends to be distributed towards the end of the bond lifetime.
Here is the math of how a 1% increase in rates would lead to trillions in losses.
Combining a duration estimate of 5.6 years with a total notional exposure of $17trn, and current Dollar price of bonds of $105.6, indicates that, to first order, a 100bp shock to interest rates would translate into a $1trn market value loss. That is using the more conservative estimate of the bond market. Using the broader bond market sizing of $40trn, the market value loss estimate would be $2.4 trillion. While Goldman believes that using the larger number "would likely be an overstatement, as much of the extra debt in the broader universe is either short maturity or floating rate", even the smaller $1 trillion loss estimate is large: over 50% larger than the market value lost in the 1994 bond market selloff in inflation-adjusted terms, and larger than the cumulative credit losses experienced to date in the non-agency residential mortgage backed securities market.
As Goldman concludes, "even if there is not a large net social loss from a rise in rates, the $1 trillion gross loss estimate suggests that some investor entities would likely experience significant distress. In the 1994 bond market decline, for example, losses on a mortgage derivative portfolio were a major factor contributing to the Orange County, California bankruptcy event. All in, the increase in total gross debt exposure, combined with lengthening bond durations and an arguably expensive bond market, suggest that rising yields should be on the short list of scenarios to be monitored by risk managers.
This ignores the losses that would also impact the Fed's own holdings of rates instruments: at last check the Fed's balance sheet had a DV01 of about $2.5 billion, or a $250 billion hit for every 1% increase in rates.
As Bloomberg adds, analysts and regulators have warned for months that rising rates will be painful for investors and lenders, but bond yields remain stubbornly low. Perhaps the reason for this is that "investors and lenders" realize that it is only a matter of time before the Fed understands it is trapped and as a result of these gargantuan losses that would be imposed on the financial industry, it simply can not hike rates. Alternatively, if there is indeed as much as $2.4 trillion in losses coming, then while bonds will be slammed, it is the equity that will be wiped out. And, as this market has demonstrated all too well, when equity selloffs start, the proceeds usually go right into bonds, making the bubble even bigger.
On the other hand, if the Fed - which has demonstated it is painfully clueless in these past few years - intends to push on with a rate hike despite a raging profits recession and a global economy that is one snowstorm away from contraction, then the trade is simple: take advantage of the algos' stupidity and short financials. Because far from "beneficiaries" of the Fed's tightening on the short end, as much as Jamie Dimon would disagree, it is the long end where the market's unprecedented duration risk is about to rear its ugly head if and when the Fed does try to "normalize."