Stephen Miller
The US 10-year bond yield is above 5 per cent. Apart from a fleeting period (less than a day) in October 2023, the last time that US 10-year yields were in this sort of territory was in June 2007.
But is the 10-year bond yield “high”? II’m not so sure.
In the first instance rising bond yields were a response to a toxic cocktail; relatively high and “sticky” inflation (exacerbated by high oil prices in the wake of the Iranian conflict); record “peacetime” US budget deficits and consequent record government debt levels against a backdrop of virtually full employment; declining demand for US bonds from official sources as the US takes a more adversarial tone in its geopolitical and commercial international relations; competing issuance from hyperscalers to fund massive AI related capex; and question marks around the independence of the US Federal Reserve.
Certain elements of that toxic cocktail are in abeyance (Fed independence). But others persist.
Some may argue that the bad news is more than adequately reflected in a US 10-year bond yield above 5 per cent. I’m not so sure.
Others may point to the possibility of resort to financial repression (official measures to suppress bond yields). Treasury Secretary Scott Bessent’s recent attempt looks unsuccessful but there might be more to play out on that front.
So, is it time to overweight bonds in a multi-asset portfolio?
Maybe, but I retain a healthy scepticism.
First, the toxicity of the cocktail nor its durability should not be underestimated.
Oil prices are elevated and inflation remains stubbornly “sticky”. There is the notion that higher oil prices are yet to be reflected in broader inflation measures.
The US budget deficit does not seem likely to shrink in the near future (witness President Trump’s promise of $US5000 to every US adult if the Republicans win the mid-terms).
Demand for US government bonds from official sources (mostly foreign central banks) has inevitably declined. It is clear why the central banks of potential adversaries like China will continue to reduce exposure to the USD and US government bonds. However, given President Trump’s somewhat capricious treatment of allies, developed country central banks may also explore diversification of their foreign exchange reserves, a la Norway.
Hyperscalers continue to issue debt to fund the massive capex requirements associated with the AI boom.
Another element at play is that US monetary policy and, indeed, broader financial conditions may not be that restrictive.
Fed Chair Warsh has openly canvassed this question.
That implies that current bond yields may not be extraordinarily high.
Between 2008 and 2022 (the period covering from the GFC to the pandemic) US 10-year bond yields averaged around 2.4 per cent. The “real” yield (nominal yield less 12-month core CPI inflation) was close to 0.1 per cent.
That was a period of extraordinarily low yields by historical standards. Yet it is etched in the minds of a number of market participants as some benchmark of “normality”.
Between 2000 and 2007 the average US 10-year bond yield was around 4.7 per cent while the real yield was around 2.5 per cent. These were not that different from averages during the 1960s (4.7 per cent and 2.2 per cent respectively).
The latter are arguably a better benchmark than that which prevailed between the GFC and the end of the pandemic.
The current figures are close to 5.1 per cent and 2.6 per cent respectively. That is in the ballpark of periods outside that book-ended by the GFC and the Pandemic.
Current nominal and real 10-year yields are probably a tad higher than long-run “steady state” nominal and real GDP growth rates. In other words, compared with r* (the long run “steady state” 10-year bond yield), the current 10-year bond yield is maybe a little on the high side. A substantial productivity dividend from AI would likely increase r* making current levels more “normal”.
And there is that the aforementioned toxic cocktail.
There are other considerations that attach to bonds in the context of a multi-asset portfolio.
It has become clear that any assumed negative return correlation between equities and bonds is highly contingent on a low and stable inflation rate. Such an environment gives central banks the wherewithal to address the consequences of a downdraft in equity markets with an aggressive reduction in the policy rate and attendant lower bond yields.
The higher inflation environment that emerged from the pandemic has rendered the negative return equity/bond correlation assumption as no longer useful.
Investors need to think about augmenting traditional multi-asset portfolios with strategic allocations to asset classes that are not as correlated with equity returns and which have some inflation protection qualities.
Certainly, infrastructure and commodities are appealing in this regard. Not only are they less correlated to equity and bond returns but they have desirable inflation protection qualities (although the Australian equity market may exhibit some positive correlation with certain commodities). Trend following quantitative portfolios also fill a similar role.
So, a 10-year bond yield above 5 per cent may well be the new normal. Macro conditions remain problematic (“sticky inflation, oil prices, US budget deficit, hyperscaler issuance etc.); bond yields are not particularly high versus historical benchmarks and compared to r*; and bonds are not as effective a diversifier to equities as they may have once been.
RBA: need to go higher
At the time, I described the August decision by the Reserve Bank of Australia (RBA) Monetary Policy Board (MPB) to leave the policy rate unchanged as “defensible but contestable”.
In essence the Board at the August meeting, ceded the argument that increases in the policy rate so far this year are working to contain inflation, and with activity growth tepid, the balance of risks pointed to no need to increase the policy rate at the August meeting.
That the decision was not without risk was exemplified by the reality that inflation pressures in Australia are among the highest in the developed world. That reflects the uncomfortable circumstance of a homegrown structural inflation proclivity.
That some Board members were cognisant of those homegrown inflation risks seemed to be made clear in the minutes of the August meeting which canvassed the potential requirement to increase the policy rate should upside inflation risks materialise.
The July consumer price index (CPI) report brought those upside risks into stark relief.
Indeed, the July report makes it is hard to construct a narrative around declining inflation. The annual rate of increase is stuck at 3.6 per cent. Perhaps even more worrying the annualised 6-monthly rate of increase is at 3.9 per cent. That is the highest in the (admittedly short history) of the published monthly series.
The July report led RBA governor Michele Bullock to tell a parliamentary inquiry last week that the risks of inflation continuing to rise were “materialising”.
What is more, “sticky” inflation must cast some doubt on the prevailing RBA (and financial market consensus) narrative that monetary policy is restrictive. A “real” policy rate of somewhere around ½ – ¾ per cent does not strike me as particularly restrictive.
I have noted in the past that Australia’s poor relative inflation performance stems from abject productivity growth which makes the task of inflation containment all the harder, necessitating higher policy rates. That abject productivity growth reflects, inter alia, the interplay of regulatory creep in labour and goods markets. Regulatory creep also imposes costs on businesses, part of which are passed on to consumers, giving further impetus to price pressures.
In the absence of a meaningful deterioration in the labour market, I strongly suspect that the RBA will be required to raise the policy rate again at its September meeting.
Moreover, at this stage meetings beyond September must be considered “live”, even if the RBA were to raise the policy rate in September.
That may be the “least worse” path even in the event of a cooling labour market.
In the wake of the oil shocks of the 1970s, developed country central banks’ big mistake was a premature retreat from an inflation focus. That “let the inflation genie out of the bottle” and resulted in the painful, but necessarily harsh Volcker medicine of the late 1970s / early 1980s.
I am not suggesting that the current circumstance is one that is anywhere near quantitatively on a par with 1970s. But there are elements that are redolent of that time, albeit on a substantially reduced scale.
The RBA’s mandate has both an inflation and employment objective. Yet the RBA has only one instrument – monetary policy – at its disposal. Trying to target two variables with one instrument is nigh on impossible as any economics undergraduate with even cursory knowledge of the Tinbergen Rule will tell you.
In his Jackson Hole remarks, Warsh noted that the Fed also has a mandate for maximum employment. He added that achieving both sides of that mandate does not mean that monetary policy works at cross-purposes since high inflation itself is very harmful to economic prosperity.
Maybe the best the RBA can do to ensure maximum employment in the medium-term is to ensure a low and stable rate of inflation.
And that is why it must increase the policy rate at the end of the month and maybe beyond.
By Stephen Miller, investment strategist
