Goldman's Francesco Garzarelli throws some numbers at its Bond Sudoku model, spins around its Wavefront Growth equity basket, and the magic firm's 8-ball spits out the following: "we presently show a 3.25% level in US 10-yr rates at the end of Q1:11. In light of the strength of the data, this now looks too low, and we would now lift the forecast to 3.5%. Our end-2011 and end-2012 projections are 3.8% and 4.3%, respectively, and we stick to these." In other words, if the market moves, we will adjust our "forecasts" accordingly. If China hike 3 more times as is expected and the 10 Year falls off a cliff, well then, we will no longer "stick to those."
From Goldman Sachs' Francesco Garzarelli
US Rates: Where Next?
The strong ISM and payroll reports last week have led to a sharp sell-off in US rates. 10-yr Treasury yields, currently at 3.63%, have broken above the 2.30-2.50% range they had been trading in since the middle part of December. In the process, the US yield curve (measured as difference between 2- and 10-yr maturity rates), has steepened, confounding the flattening priced into the forwards. We have been arguing since last November that the trajectory for rates was pointing upwards, and have recently played the market from the short side in Euroland. But the speed of adjustment in US fixed income continues to be faster than what we envisaged. Where does all this leave us?
Starting with valuations, our Bond Sudoku model (which takes into account macro expectations in the US and overseas at a 1-yr horizon) indicates that 10-yr US rates are now around 1 standard deviation ‘cheap’ relative to ‘fair value’ (3.3%). This statement presupposes that the macro outlook is stable, which is probably too strong an assumption. Our US team, for example, has shaved 25bp from their unemployment rate forecasts after last Friday’s numbers and has signaled some upside risks to core inflation. Similarly, the US Wavefront Growth equity basket, which helps us infer changes in growth expectations priced in the S&P500, indicates that an extra 9bp have been added to US real GDP growth over the past week alone.
These considerations pertain to changes in the output gap, which now appears to be closing somewhat faster than previously assumed. If this were a ‘normal’ economic upswing, it would probably be all that matters. But the financial recession of 2007-09 has opened up a lot of slack in the advanced economies and left an overhang of debt. As Fed Chairman Bernanke reminded us last week, the level of the unemployment rate (9%, well above the 5.5% ‘official’ estimate of NAIRU) and that of core inflation (60bp, the lowest since the 1960s) both remain very far from the central bank’s medium term objectives of full employment and price stability. Indeed the longer-term risk to asset prices is not the modest upward drift in rates in response to better growth news, but would arguably be if key macroeconomic variables were not to show signs of mean-reversion after all the policy stimulus has been delivered.
To emphasize this point, and illustrate the implications for yields, in last October’s issue of our Fixed Income Monthly we ran “Taylor rule’ type regressions (which link yields to the unemployment gap and the distance between core inflation and the central bank’s target) for US policy rates and for different maturities along the term structure. Repeating that exercise today, we find that ‘optimal’ yields from overnight to the 3-yr maturity should still be negative, 5-yr notes should trade around 1.5% and 10-yr and 30-yr yields trade around 3% and 4%, respectively. Admittedly the standard deviations around these estimates are very large (in the region of 100bp), but the market has now factored in a fair amount of ‘cushion’.
Based on these considerations, the ‘bond debate’ can be currently couched in terms of a trade-off between the level and changes of the output gap. The former indicates that the front-end should remain well-anchored until sufficient slack has been removed, while the latter will signal how fast the adjustment is taking place. For now, our conclusions are as follows:
Simplifying, one can think of expected government bond returns as determined by three macro factors: future growth, inflation and policy rates. Given the large amount of ‘slack’ in the US and other large advanced economies, we are of the view that underlying inflationary pressures and policy rates will remain dormant at least the next 6-9-months, even accounting for the pressures on commodity markets. The Fed’s asset purchase program is providing an additional ‘anchor’ to the yield curve in the 3-5-yr sector.
The dominant driver for (real) bond yields in the next couple of quarters will be the interplay between growth expectations and their realizations (in an upward trend). In this respect it is noteworthy that, since last October, the 2-10 slope of the US yield curve has tracked closely the US Growth Wavefront equity basket. It is also important to note that two-thirds of the re-pricing of bond yields since the ISM report has taken place from the real term structure. We calculate that 5-yr real rates, 5-yr forward (using the swaps market) are now hovering at 2%, where they stood in 2004-05 (readers who have attended our strategy conferences in Europe may recall that the title of our presentation on fixed income is ‘It’s for Real’).
If what we said in the previous point holds, it’s too early to think that bonds will pull the rug from under the equity market’s rally. Particularly since the level of bond yields across the US curve is still not restrictive (both our valuation model, and the ad hoc Taylor rule exercise mentioned above indicate that rates are at the upper end of a normal ‘valuation range’). On our baseline case, bond and equity return correlations will be low or negative – bonds selling off when stocks rally, and vice versa. What would start taking the shine off the stock markets would be an extension of the sell-off to the 4.0-4.25% area, especially if quick or not supported by a corresponding upward revision of macro expectations. This could be triggered, for example, by an unwarranted rise in longer-dated inflation expectations, which would hurt bonds and equities alike. We do not expect this and the data doesn’t suggest its coming, but we are definitely watching the Michigan survey data and inflation breakevens closely.
Turning to our forecasts and trades, we presently show a 3.25% level in US 10-yr rates at the end of Q1:11. In light of the strength of the data, this now looks too low, and we would now lift the forecast to 3.5%. Our end-2011 and end-2012 projections are 3.8% and 4.3%, respectively, and we stick to these. We will continue to look for opportunities to trade global fixed income from the short side, when opportunities arise. These will most likely continue to occur outside the US. For those readers looking for macro ‘hedges’ to an upbeat outlook for pro-cyclical assets in the advanced economies, we contend that intermediate maturity government yields fit the bill.
Where rates are still unreasonably expansionary is in large parts of the so-called growth markets, particularly Eastern Europe and Asia. Here forward rates do not sufficiently discount the inflation potential in these economies, where the output gap has closed, and credit growth is too brisk. In EMEA space, Poland and Turkey stand out. In Asia, as Fiona Lake has argued in her inaugural issue of the Asian Interest Rate Monthly, inflation risks are most pronounced in China, India, Indonesia and Korea, leaving the rates markets in these countries most vulnerable. While higher US rates would impact most markets in the region, Hong Kong and Singapore are most directly exposed to a rapid sell off in US rates