Strategy commentary cont'd: Government bond yields finished sharply higher. The 10-year US Treasury yield rose 55bps to 5.28%, leaving it 113bps above its level a year earlier. US 10-year and 30-year yields reached their highest levels since 2007 and 2002 respectively. German 10-year Bund yields climbed 26bps to 3.59%, having breached 3.60% during the month, while 10-year UK Gilt yields rose 36bps to 5.42%. French OATs were the notable underperformer: 10-year yields surged 68bps to 4.86% and 30-year yields traded above 5% for the first time since 2008. Italian 10-year BTP yields rose 46bps to 4.61%. The OAT/Bund spread widened 42bps to 127bps and the BTP/Bund spread 20bps to 103bps. Japanese 10-year government bond yields increased 11bps to 3.05%, up 141bps over the year.

Oil was the clearest expression of the geopolitical risk premium. Brent crude rose 14.4% to US$103.53 a barrel and WTI 5.4% to US$90.42, leaving the benchmarks 54% and 45% higher than a year earlier. Non-interest-bearing precious metals moved the other way as yields climbed: gold fell 6.3% to US$4,157 an ounce and silver declined 9.2%, unwinding much of August's rally, although gold remained 7.7% higher over the year. Bitcoin gained 6.05% to US$83,621 but remained 27.1% below its level a year earlier.

The US dollar strengthened against the euro despite the ECB's hike, with EUR/USD falling 2.5% to 1.133. The yen recovered some ground following the Bank of Japan's tightening, with USD/JPY declining 1.5% to 157.41.

Equity gains were highly concentrated. The S&P 500 fell 0.45% in price terms and returned -0.35% including dividends, with technology the only sector to finish the month higher. The Nasdaq 100 gained 3.2%, or 3.3% including dividends, supported by semiconductors. Elsewhere, the Euro Stoxx 50 fell 2.35% (-2.2% total return), the Euro Stoxx Banks Index declined 1.7% (-1.4% total return), and the FTSE 100 lost 2.0% on both measures after rounding. The Nikkei 225 rose 0.7%, or 1.2% including dividends.

Credit and primary issuance

Credit spreads widened alongside the rise in government bond yields. US CDX IG widened 3bps to 60bps and CDX HY 27bps to 341bps. In Europe, iTraxx Main widened 6bps to 63bps, Xover 15bps to 301bps and senior financials 8.5bps to 69bps. Despite September's widening, Xover remained 27bps tighter over the year and CDX HY 5bps tighter.

Cash credit also weakened, particularly in Europe. Euro aggregate corporate spreads widened 11bps to 90bps as the post-summer supply surge repriced secondary markets. Sterling corporate spreads widened 5bps to 91bps and US investment-grade corporate spreads 3bps to 80bps. Interest-rate exposure remained a major driver of corporate bond returns: the Global Aggregate Corporate Index hedged to US dollars lost 2.10% in September, while its duration-hedged counterpart gained 0.30%. Over the year, these indices returned -0.96% and 4.74% respectively. The 5.7 percentage-point gap illustrates how heavily duration has weighed on returns in this cycle.

US primary market activity was robust, with US$210bn of US dollar investment-grade supply, of which corporates accounted for 52%, according to Bank of America. Non-US banks led issuance following the Labor Day break. Despite the heavy calendar, new-issue concessions remained contained at around 2.5bps on average. Citi's US$12bn four-tranche holding-company deal attracted US$40.5bn of demand, with the long end tightening 3bps immediately after pricing. Acquisition financing was an important driver of corporate supply, including Aon's US$13.5bn seven-tranche deal to fund its US$17bn purchase of USI.

European issuance also accelerated as investors returned from the summer break. According to Barclays, €107bn of investment-grade supply priced, including €47bn from financials. Average new-issue concessions across financials were 6bps, compared with around 2.5bps across the US dollar market. Although some new bonds tightened after pricing, several issuers' outstanding bonds repriced wider towards new-issue levels. UBS's €1.75bn 11NC10 holding-company transaction offered an all-in yield above 4.5%, attracting strong demand from insurers. Books were 3.7 times covered and the bonds tightened 4bps immediately after pricing. Amazon issued a four-tranche £4bn deal against £10.7bn of demand, with the shorter-dated tranches tightening 2-3bps in secondary trading.

Strategy commentary cont'd: Australian primary markets were quieter than August's frenetic pace but remained ahead of last year. Approximately A$10bn of investment-grade credit was issued, 9% more than September 2025's A$9.2bn. Year-to-date supply reached A$154.6bn, compared with A$120.4bn at the same point last year. Issuers navigated greater credit-spread volatility during a month in which both the RBA and the Fed raised rates by 25bps against a backdrop of elevated oil prices and other inflationary pressures.

Securitisation was a significant contributor. ANZ issued A$2.5bn in the second instalment of its Kingfisher RMBS programme. The senior A1 tranche attracted A$3.4bn of demand for an A$1.9bn issue, or 1.8 times coverage. Bendigo and Adelaide Bank's first Torrens RMBS transaction of the year raised A$920m against A$1.38bn of orders. Great Southern Bank, formerly known as Credit Union Australia, issued A$500m of five-year senior debt against A$900m of demand.

Subordinated supply also picked up. Macquarie Bank issued A$1.4bn across 10.5NC5.5 fixed- and floating-rate tranches and a 15NC10 floating-rate tranche, attracting A$3.3bn of demand. Transgrid followed with a 10-year fixed-rate senior bond and a 30NC5 floating-rate subordinated note, together raising A$950m against A$2.8bn of orders.

Nationwide's five-year floating-rate covered bond and ING's five-year fixed- and floating-rate covered bonds together raised A$2.5bn against A$4.2bn of demand. Kangaroo issuance continued through Mizuho, Anglian Water and E.ON, which priced senior bonds across three-, five- and six-year tenors for a combined A$1.65bn against A$2.9bn of orders. Semi-government syndicated issuance added A$5.2bn through NSWTC, WATC, NTTC and TASCORP.

Australia and New Zealand

Australian government bonds sold off again, although their 10-year yield rose less than the US equivalent. The 10-year Commonwealth government bond yield increased 26bps to 5.35%. Domestic credit spreads were mixed: five-year major-bank senior spreads widened 2bps to 67bps and subordinated spreads 3bps to 130bps, while hybrid spreads tightened 16bps to 159bps. The Australian iTraxx Index widened 5bps to 72bps.

The rise in yields weighed on domestic fixed-rate benchmarks. The AusBond Composite Index declined 0.86% and the AusBond Credit Index lost 0.54%, while the AusBond Credit FRN Index gained 0.37%. The Australian dollar fell 3.1% against the US dollar to US$0.6947. Australian equities underperformed the major overseas indices discussed above, with the ASX200 falling 3.2% in price terms and 2.4% including dividends.

New Zealand's 10-year government bond yield ended September at 5.04%, after reaching 5.13% on 28 September. The New Zealand dollar fell 4.8% against the US dollar to US$0.5634, while the NZX50 price and total-return indices declined 1.5% and 0.6% respectively.

The RBA raised the cash rate by 25bps to 4.60% at its September meeting, moving closer to the 4.75%-5.00% range implied by our Taylor Rule modelling.

Portfolio performance

September was challenging for fixed-income investors, with widening credit spreads adding to the pressure from higher government bond yields. The daily liquidity and AAA rated Coolabah Active Sovereign Bond Fund's Zero Duration Class returned 0.38% net of fees (0.44% gross), compared with 0.35% for its RBA Overnight Cash Rate benchmark and the 0.88% loss suffered by the fixed-rate AusBond Treasury Index.

The Floating-Rate High Yield Fund, with an average A+ portfolio credit rating and zero interest-rate duration, returned 0.31% net of fees (0.38% gross) for its Institutional Class. While most floating-rate strategies generated positive returns, fixed-rate strategies were weighed down by the sharp rise in yields.

Strategy commentary cont'd: For investors considering new allocations, the repricing has improved the opportunity set. A cash rate of 4.60% and a 10-year Commonwealth yield of 5.35% are sitting way above long-term historical levels. We continue to favour averaging into fixed-rate exposure rather than trying to identify the precise peak in yields. If the housing correction eventually tips Australia into recession, the RBA could cut well below neutral, and high-quality fixed-rate government bonds could be major beneficiaries.

The return arithmetic is increasingly attractive. Consider an illustrative 10-year government bond bought at par with a 5.35% annual coupon, paid semi-annually, and a 5.35% yield. Its modified duration is approximately 7.7 years. After 12 months, when the bond has nine years remaining, a rise in its yield to 6.35% would produce a capital loss of about 6.8%. Adding the year's 5.35% coupon income gives a total return of approximately -1.4%. If the yield instead fell to 4.35%, the capital gain would be about 7.4%, producing a total return of approximately 12.7%.

Under these assumptions, the bond's yield would need to rise approximately 78bps over the year before the investment lost money. These are illustrative absolute returns, before fees and tax and without coupon reinvestment, rather than forecasts or returns relative to cash. The income cushion does most of the work, while convexity adds a further benefit: an equal-sized fall in yields produces a larger capital gain than the loss caused by a rise. Persistent inflation and higher term premia remain risks, but we believe the combination of starting income and potential capital gains supports progressively building fixed-rate exposure.

At 30 September, the Floating-Rate High Yield Fund offered a gross running yield of 7.1% a year. The Long Short Opportunities Fund and Global Carbon Leaders Fund (CBNX) each offered 7.0%; the Global Floating-Rate High Yield Fund (YLDX) and Long Short Credit Fund offered 6.8%; and the Active Global Bond Fund (GFXD) offered 6.6%. Running yields are not forecasts of total returns, which also reflect changes in asset prices, hedging and fees.

Please note that past performance is not a reliable indicator of future performance. Investors should read the relevant Product Disclosure Statement and Target Market Determination before making any investment decision and consider obtaining advice from an independent financial adviser to determine whether an investment is appropriate for their objectives, financial situation and needs.

The AI boom could end in a nasty global recession

As home values fall at a double-digit annualised pace, the RBA is doubling down on higher rates, setting the stage for what could be the worst housing correction in Australia's history. Markets, meanwhile, continue to under-price the scale of the global tightening cycle that the AI investment boom is helping to drive.

Strategy commentary cont'd: The RBA appears unmoved by the first-order consequences of the housing downturn. Home values across the five largest capitals have fallen 3.9% over the past three months, a 15% annualised decline. Sydney has lost 4.6% in a quarter, or 17% annualised, Melbourne is shrinking at a 13.5% annual pace, and even the erstwhile boom markets of Perth, Brisbane and Adelaide, which still show positive year-on-year gains, are now falling at double-digit annualised rates.

Governor Michele Bullock has described the downturn as not "entirely out of line with the sorts of downturns we've seen in the past", sitting "in the lower range" of prior corrections and "partly monetary policy working", while noting that prices remain about 50% above early 2020 levels. The implication is that values have a long way to fall before anyone at Martin Place loses sleep.

And fall they will. The RBA lifted the cash rate to 4.60% in September and several economists expect a further move to 4.85% by year-end, which would be the highest level since late 2008. If the cash rate ultimately heads towards 5%, Australia is likely to experience the largest housing correction in its history.

What is striking is the scale of the RBA's capitulation. Since inflation targeting began in 1993, the bank's five hiking cycles have averaged 265bps, with a median of 275bps and a smallest of 150bps. In the last cycle it deliberately lagged the world, stopping at 4.35% while the Fed, Bank of England, Bank of Canada and RBNZ went to between 5% and 5.5%. This was the "narrow path": hike as little as possible, tolerate above-target inflation for more than half a decade, and hope to preserve post-COVID employment gains that rested on extraordinary fiscal deficits, the lowest interest rates in history and rapid population growth.

Earlier this year the RBA signalled it might be one and done, or at most reverse last year's 75bps of cuts. Having now done exactly that, it has become markedly more hawkish. In parliamentary testimony, Bullock said the upside risks flagged in August "appear to be materialising", that "inflation is too high", and that the question is "whether the tightening in monetary policy to date will be sufficient to bring inflation back to target in a reasonable time". Deputy governor Andrew Hauser went further, warning that the bank "will do what it takes" and that the difficulty of the task does not mean "3, 3.5 or 4 [per cent] is good enough".

Three shifts underpin the change. First, the RBA's models now put the 4.35% cash rate around the top of the range of estimates of nominal neutral, so policy was never as restrictive as the doves assumed; Bullock cites an AI investment boom that is putting "pressure on the savings pool", stimulatory fiscal policy, remilitarisation and the onshoring of supply chains as structural forces lifting neutral. Second, the bank concedes that the lowest unemployment rate consistent with stable inflation "might be higher for a while". Third, the "immaculate disinflation" thesis, under which inflation could return to target without materially lifting unemployment, is dead and buried: if restoring price stability requires "a bit more slack in the economy than we thought", then so be it. Martin Place, in Bullock's words, "retired the narrow path a while ago".

Hauser has also endorsed Kenneth Rogoff's argument that the recent bond sell-off is not the aberration; the anomaly was the era of near-zero rates after the GFC and during COVID, and long-term real yields are now "rather closer ... to a more sensible level". That forces a reassessment of where the cost of capital settles, and markets habitually struggle to price a full hiking cycle. The Fed's typical cycle since 1994 has run to roughly 300 to 330bps, with the smallest, in 1999-2000, at 175bps. Yet following the Fed's 25bp rise in September, its first since 2023, markets were pricing only about 90bps of further tightening, with the funds rate peaking near 4.8% in September 2027. Our Taylor rule modelling implies a total of 125bps of Fed hikes assuming a 3.6% neutral rate, and 150 to 200bps is conceivable if fiscal policy remains highly expansionary. Bond markets have been too sanguine for too long, and equity markets are more complacent still.

Herein lies the real danger in the AI story. The capex surge is a conventional investment boom arriving at precisely the wrong moment, adding a demand shock to an economy already running hot amid enormous public spending. It could ultimately cannibalise itself by forcing risk-free rates much higher, crushing the animal spirits, demand, investment and growth it initially unleashed. It could all end in a nasty recession, much like its predecessor, the great tech boom of the late 1990s, which culminated in a circa 49% decline in equity valuations. It is not clear that markets have fully figured this out.

Strategy commentary cont'd: Kevin Warsh the hyper-hawk has only just begun

A conventional Fed hiking cycle would be brutal for equities, real estate and private markets. Yet investors are behaving as though the first Fed hike in three years is good news.

Kevin Warsh has revealed himself to be a hyper-hawk in response to a belligerent inflation cycle fuelled by the hyperscalers' unprecedented investment programs. The bad news for asset prices is that the typical Fed hiking cycle involves about 300bps or more of rate rises, whereas markets are pricing only around 100bps.

The Fed has been well behind the curve, so it was encouraging to see its new chairman deliver, lifting the federal funds target range by 25bps to 3.75% to 4.00% in the first rise since 2023, on a unanimous vote. The Fed's starting point of about 3.6% was almost identical to the RBA's cash rate when it began raising this year, and the Fed is late to the party, following the RBA, the ECB, the RBNZ and Norges Bank, among others. Since last year Coolabah has argued the Fed would need to lift rates by 75 to 100bps. Our current modelling implies a total of about 125bps is required, and we suspect Warsh may ultimately need 150 to 200bps. The opening salvo pushes the world further into our long-held central case of a globally synchronised tightening cycle.

The most telling insight into where the Fed sits in the cycle was Warsh's description of the first move as having merely "removed a dose of accommodation". Policy, in his telling, is still stimulatory, so the target rate sits below neutral rather than above it, as the Fed had claimed for years. The Fed has gradually lifted its median estimate of neutral from 2.5% after the pandemic to 3.2% in its September projections, and two of the 18 participants now put it at 3.875%, lending weight to Warsh's view that one would be "hard-pressed to describe broad financial conditions as restrictive". With unemployment at just 4.1% and inflation well above target, his argument is that the policy rate is plainly not doing enough of the heavy lifting. It is a change of tune for a chair who was avowedly dovish before his appointment, but it represents a capitulation to the global central banking zeitgeist that neutral rates are rising, with the corollary that policy rates need to climb to snuff out pressures that have persisted since the pandemic.

The Fed's challenge is stubbornly sticky inflation, with core PCE running at 3.5% annualised over six months and 3.3% over 12. The culprit is the collision of an ultra-low jobless rate, huge budget deficits and the mother of all capex booms powered by spending on AI infrastructure, which together generate excess demand while running into a panoply of supply chain bottlenecks. Break apart the core PCE index and consumer technology prices are surging, the first genuine bout of technology-product inflation in decades, alongside rising goods prices courtesy of the wars in Iran and Ukraine and persistent services inflation powered by a tight labour market.

Warsh was composed and resolute in his press conference, hammering home the message that the Fed is unconditionally committed to returning inflation to target in a timely fashion, with the risks skewed to the upside. Asked how the dot plots, which point to a peak funds rate of only 4.1% while leaving core inflation above target until 2029, squared with that aspiration, he retorted that "those aren't my forecasts". He also stressed that while the Fed cannot prevent supply shocks, it can stop them spreading: "What we can do, and will do, is ensure that any changes in relative prices don't broaden out, don't have second- and third-order effects in the economy." The Fed is now in the demand-destruction business. It wants to slow growth, loosen the labour market and create enough spare capacity to bring demand and supply back into equilibrium. The dot plots nonetheless remain dovish relative to our Taylor rule modelling: the median participant expects one more hike this year, no change in 2027 and cuts in 2028 and 2029, although eight of the 18 project a 4.375% funds rate by end-2027, implying another 50bps of tightening from the current midpoint.

Warsh also appears to have put considerable distance between himself and President Trump, who revealingly claimed he had told Warsh he "might as well vote with the board because it's not going to matter". Treasury Secretary Scott Bessent's attempt to suppress long-end yields through enlarged bond buybacks failed to deliver a sustained reduction in borrowing costs, with the ensuing sell-off driving the 30-year yield to about 5.37%, at the time its highest since 2007, before it pushed higher still. That policy failure arguably created more room for the Fed to hike and for Warsh to demonstrate his independence from the White House, and coupled with a brisk 0.4% rise in the August consumer price index, it made a September move increasingly difficult to avoid.

Strategy commentary cont'd: The more pressing question is whether Warsh follows up in October, six days before the midterm elections, timing many had assumed would be an impediment. He is at least 100bps behind the curve and knows that one hike will not meaningfully recalibrate the cost of capital. That requires a sequence of increases, making both October and December live prospects if the data remain strong.

How far could this cycle run? Across all Fed hiking cycles since 1955, the median increase in the policy rate has been 313bps and the average 478bps. Since 1994, the mean has been 330bps and the median 300bps, and even the shallow cycles of 1986, 1999 and 2015 delivered 137bps, 175bps and 225bps respectively. Against that history, our call of about 125bps looks modest, and even the 150 to 200bps we think investors should contemplate would constitute an unusually skinny hiking cycle.

A conventional 300bp move would be brutal for equities, real estate and private markets. Defaults would rise, insolvencies would multiply and capital would migrate towards liquid, lower-risk alternatives as attractive risk-free returns made expensive risky assets look uncompetitive. Yet equities are behaving as though the first Fed hike in three years is good news. That is a bold wager. If history is any guide, Warsh has not finished tightening, and there remains much downside in asset prices.