If at first you don't succeed, try, try, keep trying again and again.
That appears to be the mantra of Goldman's credit strategists, who one year ago when the 10Y was trading right around current levels, predicted that the yield on the benchmark bond would rise to 3% by the end of 2015. It was just a little off.
First, here is Goldman's apology for being so wrong in what may be the most important forecast a bank makes (as it involves everything from the economy to the equity risk premium): where the 10 Year will trade:
We are ending the year with the benchmark 10-year Treasury yield at around 2.3%. We started 2015 expecting yields to close the year at 3.0%, or 50bp above the forwards prevailing at the time. Over the course of 2015, we progressively lowered our year-end projection to incorporate incoming information, specifically: in March, we shaved our forecast by 50bp to 2.5% (with the year-end forwards prevailing at the time at 2.1%). By June, yields had bounced back to 2.5%. In October, following developments in China and oil markets, we brought our forecast down further to 2.3% (when the forwards were at 2.1%).
Why was Goldman wrong: "These forecast revisions mainly reflected three developments:"
- US real GDP expanded at a considerably slower pace than we expected at the beginning of 2015 (our latest growth forecasts are 2.5% for 2015 and 2.3% for 2016; the corresponding numbers at the start of 2015 were 3.3% and 3.0%).
- We forecast that the Fed would hike rates twice this year (in Q3 and Q4), a view we subsequently changed to only one hike in December.
- The decline in crude oil prices - resulting from supply-side factors and anomalous weather patterns - proved to be a larger and more persistent drag on headline inflation (and more broadly, a source of asset price turbulence) than we thought at the start of this year. This acted to depress the term premium globally (particularly in the Euro area and Japan, where inflation targeting central banks are conducting QE) and raise credit risk.
In other words, Goldman got virtually everything wrong, except where the S&P would close: they had been steadfastly predicting 2,100 for most of the year, and then lowered it to 2,000. There is still time for the S&P to tumble 70 points and hit Goldman's forecast. Indicatively, Goldman also expects the S&P to be at 2,100 on December 31, 2016 and 2,200 at the end of 2017.
* * *
Ok fine, everyone makes mistakes when predicting the future (about pretty much everything). Where things get comic is that instead of admitting it needs to shift its forecasting process, Goldman - which is expecting a dramatic drop in one-year forward GDP from 3.6% to 2.3% - is expecting that on December 31, 2016 the 10Y will yield exactly the same 3.00%.
In other words, an identical forecast from last year. Here's why according to text that may have been lifted from a year ago:
Looking ahead, we forecast 10-yr benchmark Treasury yields to rise to 3% by end 2016 - or 50bp above the forwards. This expectation is built on the following set of assumptions.
- An ongoing above-trend economic expansion: US 10-year Treasury yields currently compare with a 'fair value' estimate of 2.7% (i.e., are around 1 standard deviation too low) based on our Bond Sudoku model. Factoring in our economists' macro forecasts for above-trend GDP growth, an acceleration of inflation and higher policy rates in the US alongside macro developments abroad, our model 'fair value' estimate increases to 3.4% by end-2016. In forecasting a 3% yield by next December, we presently assume the 'valuation gap' to remain the same as today's, to reflect the ongoing QE in the Euro area and Japan.
- Fed hiking more than priced into the forwards: An important ingredient in our estimates is our US Economics team forecast of the Fed hiking four times this year, or by a cumulative 100bp to 1.30%. The market instead sees rates at 90bp after the 14 Dec 2016 FOMC meeting, with two-thirds of the probability mass in a 60-110bp range. Using the market pricing for the effective Fed Funds rate, our measure of 'fair value' for Treasuries would be around 10-15bp lower (or 3.25%), all else equal.
- An upward shift in medium-term inflation expectations: We see CPI inflation picking up during 2016, led by the services categories. Although this is already reflected in the fair value estimates discussed above, an increase in service inflation (ex-energy) from the trailing 2.9% towards 3.5% (levels last seen in 2007-08) may well act to lift medium term inflation expectations, and thus the term premium. Such an outcome underpins our view on that the 2-10-year nominal term structure will steepen over the next couple of quarters. To be sure, the high volatility and downward pressures in crude oil prices continues to keep the term premium depressed. But we expect these effects to wane and reverse during the course of 2016, as crude oil prices increase towards our 12-month forecast of US$ 50 per barrel on WTI, from US$ 37 per barrel at the time of writing.
- A higher expected terminal rate and an increase in term premium: An historical analysis of the behaviour of the US yield curve during previous four Fed tightening cycles reveals that the market has always progressively revised upwards its view on the 'neutral' level of rates as the Fed pushed policy rates upwards (see 'How Will Bond Yields Move in the 100 Days After the First Hike ', 11 Sep 2015). In other words, expectations on the degree of monetary tightening the economy could withstand have in the past tended to be adaptive. We judge the current market-implied level of neutral rates (around 50bp in real terms) as too low. By contrast, as the hiking cycle extends, the norm is for the 'term premium' to decline from relatively high levels in proportion to terminal rate expectations (on the average of the past 4 cycles, around 75%), a phenomenon particularly evident in the 2004 cycle (here we use the NY Fed's ACM estimate of the term premium). Presently, the term premium is zero. Although there have been periods, like in the 1960s, when the term premium was very low, it has not been negative outside short periods of time.
Since we expect all of these predictions to be wrong again (check back in a year's time to find out why), it is more prudent to read Goldman's own hedging risk factors:
Outside changes in macro developments, we see two main big risks to our Treasury yield forecasts: A further decline in oil and commodity prices, leading to broader asset price turbulence, would most likely lead to lower bond yields than we presently forecast by amplifying deflationary expectations. At the other end of the spectrum, an increase in yields (and their volatility) in the Euro area and Japan, spilling over into the US market. Our analytics indicate that the decline in both JGBs and Bunds has contributed to keeping US Treasury yields lower than they would otherwise have been this year.
As for the punchline, recall that Goldman's top trade for 2016 is to be long the USD vs the EUR and JPY:
Top Trade #1: Long USD vs short EUR and JPY
Go long USD against an equally-weighted basket of EUR and JPY at 100, with a spot target of 110 and a stop loss of 95. Annual carry is positive at around 1%.
Well, it is not even 2016 yet, and Goldman, according to its latest trade update, is already half way to being stopped out.
Stay long USD against an equally weighted basket of EUR and JPY, opened on 19 November 2015 at 100, with a spot target of 110 and a stop loss of 95, currently trading at 97.68.
Good luck to all.