It all started in February, when we first reported that something unexpected had happened: for reasons that were at the time unknown, the global credit impulse had unexpectedly tumbled, turning negative, a move which we predicted would result in a steep slide in the “soft” economic data, end the “reflation” optimism and unleash a wave of dovishness from the Fed.
Then, two months later when the reflation trade was officially over, in early April the culprit for this sudden collapse in global growth momentum was identified: China, which together with the price of oil, had been the only catalyst for the global reflation trade since the “Shanghai Accord” in February 2016, and had seen its credit impulse crash at the fastest pace since the financial crisis, dropping to a level not seen since 2010.
Fast forward to this weekend, when in his latest fixed income weekly report, Deutsche Bank’s Dominic Konstam – who two weeks ago flipped back from being passionately bearish on rates and TSYs, and a supporter of the reflation trade, to anticipating much lower yields and a slowdown in the economy – writes that “It was good while it lasted.“
What he is referring to is both the so-called “bear market” in bonds as well as global growth momentum, and in taking a page from the second derivative playbook we have focused on over the past two months, he expects it to decline sharply now that US excess liquidity has reached a downward inflection point, and as a result for the foreseeable future monetary policy will be a headwind: “historically, the peak in G3 yields has coincided with peaks in output momentum. It would not be unreasonable to expect at least a 50 bps decline from peak to trough in G3 yields following a peak in yield momentum.”
Below, he explains why his latest advice to rates traders is to “Sit back, relax, the bear market in bonds may be over.”
G3 yield momentum has swung from a monthly trough of -2 percent in early 2015 to the recent peak of almost +60 bps in March, 2017 using 5y5y government bond yields for France, US and Japan (equal weights France, US and 15 percent for Japan). Momentum is the change in the year versus the previous year. At current levels momentum is still strong at +38 bps. But if yields stay where they are momentum will rise to new peaks of +82 bps in June. Yet according to global output momentum (change in year over year output growth), yield momentum itself should be going in the opposite direction. Even if momentum was to fall to zero, a “neutral” zone for 5y5y global yields would need to be 70 bps lower by June versus current levels; if the market was more patient, 50 bps lower by October, which was before the accelerated sell off around the US election. Flat momentum wouldn’t be reached until December, 2017.
The concern for excessively bearish yield momentum reflects the fact that when output momentum peaks or troughs, there is a strong correlation with troughs or peaks in yield momentum with varying degrees of subsequent adjustment depending on the nature of the shift in output. The correlation isn’t always contemporaneous but it is always within a few months either side. One can’t help but think that if output momentum is rolling over, then we have already seen a clear signal of the rollover in yield momentum and if that is the case the recent (local) bear market is over and the only question is the extent to which it can unwind.
Ironically the bear market, since it was never that strong, failed to break the long run decline in global yields. If the bear market in bonds is dead, the bull market may never have noticed.
For those who wish to recreate their own excess liquidity tracker at home, which together with Chinese credit impulse are arguably the two most important leading indicators in a time when credit creation is the only thing that matters, here are some further details.
Our starting point for the shift in output momentum is the excess liquidity model. It is very clear that there is pending a decent loss of momentum predicted by the rollover of excess liquidity. This would also serve to reconcile the hard and soft data debate.
For those investors less familiar with the excess liquidity model, it is simply the difference in the first derivative from a Fisher equation, MV=PT where V is considered constant. The only adjustment we have made historically to V is during financial crisis when it was marked lower by 20 percent. M is very broadly defined to include MZM and all credit outstanding adjusted for central bank and bank holdings of debt to avoid double counting. PT represents nominal output and for monthly data we use finished goods producer prices, excluding oil to reflate industrial output. Excess liquidity itself is highly correlated with output momentum, regardless of whether it is reflated with a one year lead. This isn’t necessarily surprising since it implies that today’s output growth is a function of last year’s credit growth and a lagged dependent variable, the latter normally ensures a decent forecasting capability reflecting the strength of the cycle. Intuitively the model works well as credit expansion is required to sustain faster growth (it “accommodates” economic growth). This may occur endogenously (as Keynes argued, real growth solicits stronger credit) but not necessarily – sometimes if credit is constrained any acceleration in real growth or pricing may simply cannibalize itself. In the latest downturn in excess liquidity the run up in producer prices has not been accommodated by faster credit expansion – credit has actually slowed slightly and hence there is deficient excess liquidity to sustain output momentum.
Finally, based on both empirical data and concurrent indicators, the correlation between turning points in output momentum and yields is well established. The charts below show various metrics for G3 yields, US yields and 5y5y yields versus both nominal and real output momentum.
There are about 15 episodes going back to 1992 which are all coincident with the shift in direction of yield momentum. For G3 yields, there are 59 months where yields are at least one standard deviation higher than a year ago and 39 months when yields are at least 2 standard deviation higher than a year ago. In the former case the subsequent move in yields averages -56 basis points and in the latter case, -61 basis points. The last major shift in momentum was 2014 after taper tantrum. In the latest episode there has been a 1 sd increase in yield momentum for March-April, 2017. It would not therefore be unreasonable to expect at least a 50 bps decline from peak to trough in G3 yields – if Japan doesn’t move, then depending on the move in European yields, this would be disproportionately larger for the US.
In short: the gamble launched with the Shanghai Accord in February 2016 which prevented a global bear market, namely the global reflation experiment is once again dead – thank China and central bankers who had hoped to hand over monetary policy to Trump’s failed fiscal policy. The market just doesn’t know it quite yet.