Robin Marshall has responded to Peter Warburton’s article titled ‘What does the flattening yield curve mean’.
The first paragraph ends with a perfectly reasonable question about the information content of a flattening yield curve and whether it signals the end of the US/UK expansion. However, the final question about buying bargain basement long-dated bonds seems mis-specified, since the flatness of the curve means an investor gets very little extra yield from buying long-dated bonds (eg, current US 30 year yield is about 3%, but the 10 year yields 2.85%, so pick-up is only 15bp for an extra 20 years of inflation and duration risk) ? It’s more likely to be a “ golden moment” to acquire 7-10 year bonds in that regard.
The problem with relying on official National accounts data for evidence of economic slowdown is that this data is a lagging indicator. Central bankers often refer to the problem of using official National accounts data as driving using the rear view mirror. This is why markets latch onto financial indicators with a strong track record as leading indicators of recessions/slowdowns, rather than lagging indicators, like the 10 yr versus 2 yr yield. A good recent example was the 2008/09 GFC, when the bond market started to rally strongly well in advance of the economic data collapsing. I accept that the information content of a flatter yield curve may be different in this cycle (a point I made in my article) because of the Fed’s holdings of MBS, which reduces the need for active duration risk management of MBS portfolios, and convexity hedging, since the Fed is not a profit-maximising private sector body. However, flatter yield curves have been a feature since the early 2000s, when Chairman Greenspan referred to what he saw as “ a conundrum” ; the conundrum has become the norm.
Also, on real rates, real yields on Tips are already +0.75% to 0.905% (5 yrs to 30 yrs), so they are comfortably positive – it is not correct to say they have not revisited positive territory, is it ? Appealing to evidence from previous post-war US tightening cycles is also a little odd. These were cycles with much higher inflation rates, whereby the so-called term premium rose steadily from the 1960s to the 1980s, as investors sought much greater inflation protection for holding bonds with fixed nominal coupons. The term premium has been falling with inflation since the 1980s, ever since Fed Chairman Volcker chose to squeeze inflation hard by tightening monetary policy aggressively.
In addition, the US Treasury didn’t issue Tips until the late-1990s, with market-determined real yields, and was very keen when they did to ensure they successfully sold their early Tips issues by pricing them very attractively, at high real yields, around 4%. So it is hard to draw comparisons with earlier cycles, particularly as before Volcker in the US, and PM Thatcher in the UK, there was virtually no attempt to target lower inflation rates, but a revealed preference amongst policy-makers to target full employment. If you look at very long time series on interest rates, going back decades, or even centuries, the biggest aberration was how high rates went in the 1970s & 1980s. It’s true that rates near zero are historically low, but we also had over 20 years with the UK base rate stuck at 3% in the Era of Cheap money from 1929 to 1952…..so these notions of mean-version, and normalisation, on interest rates are very sensitive to the period over which the mean is calculated ! Recent work in behavioural finance and economics – notably Kahnemann et al – has suggested investors may get anchored in the period they are familiar with, and frame expectations accordingly, about what are “ normal “ rates….