The ASX hands back its August gains as rates reprice

The ASX 200 finished August with a fifth straight month of gains, having set a record high of 9,296.7 early in the month. September has been a different story. The index broke below 9,000 in the second week, lost almost 3 per cent in a single week and hit an 11-week low of 8,657 on 15 September. It closed at 8,680 on 28 September, the day before the RBA’s decision, leaving it down a little over 4 per cent for the month. For context, September has historically been a soft month for the local market, with an average fall of around 1.8 per cent, so this one has been weaker than usual.

The main driver has been interest rates rather than anything in company earnings. When bond yields rise, the value investors place on future profits falls, and rate-sensitive parts of the market such as banks, listed property and technology feel it first. With Australian 10-year government bond yields sitting around 5.4 per cent and the RBA lifting rates on 29 September, that pressure has been hard to avoid.

August reporting season gave a mixed picture. Aggregate earnings growth returned after several flat years, carried largely by the resources sector, but company commentary was cautious, with many management teams pointing to slowing growth, persistent cost pressures and more uncertainty heading into FY27. Share price reactions had far more to do with what the market had already priced in than with the headline profit number.

On the broader economy, GDP grew 0.4 per cent in the June quarter and 2.1 per cent over the year. Household spending remained subdued outside of a surge in vehicle purchases, and imports absorbed a good share of the demand. Unemployment ticked up to 4.6 per cent in August, according to figures released on 24 September, although that came alongside a strong lift in the number of people looking for work, with participation back up to 67.1 per cent, close to its record high.

Global markets: strong economies, and bond yields to match

The defining feature of global markets this month has been a coordinated move higher in interest rates. On 16 September the US Federal Reserve raised rates by 0.25 per cent, unanimously, its first increase in more than three years, taking the federal funds range to 3.75 to 4 per cent. Chair Kevin Warsh said inflation had been too high for too long, pointing to a strong economy and ongoing tension in the Middle East. The Fed’s own projections suggest a second increase before the end of the year is more likely than not.

The Fed was not alone. The European Central Bank lifted its deposit rate to 2.50 per cent on 10 September, warning that inflation would remain well above target for an extended period, and the Bank of Japan raised its policy rate to 1.25 per cent on 18 September, the highest level in Japan since 1995. Bond markets have responded accordingly. The US 10-year yield pushed above 5 per cent mid-month, and on 23 September the 5-year yield reached 5 per cent for the first time since 2007, after business activity data came in well ahead of expectations. By 29 September the 10-year yield was sitting at around 5.2 per cent and the 30-year yield had reached its highest level since 2004.

Given all of that, US shares have held up better than you might expect. The Nasdaq closed at a record high on 22 September, and although US shares have given some ground back since, with the S&P 500 down 0.8 per cent on 28 September, the broader market remains supported by solid earnings. The catch is how narrow the leadership has become. A large share of recent gains in the Nasdaq 100 has come from a handful of chipmakers and technology names, which leaves the index more exposed to a change in sentiment around those few companies than its headline number suggests. European markets have been softer, with the German and UK indices both falling around 1.5 per cent on 18 September alone.

Oil remains the wildcard. With the Iran conflict unresolved, Brent crude has spent most of September above US$100 and was trading around US$105 a barrel on 29 September, after a brief dip below that level the week before. On trade, President Trump hosted Chinese President Xi Jinping in Washington on 24 September. The two sides agreed to extend their tariff truce by two months, to 10 January 2027, and to planned tariff reductions on around US$60 billion of two-way trade. A lasting deal is still some way off, but any easing of tension between the world’s two largest economies is good news for Australian exporters.

AI: a wobble, a rebound, and a question the RBA is now asking out loud

The most unusual market story of the month came from within the AI industry itself. In mid-September, the heads of several of the leading AI developers, including Anthropic’s Dario Amodei, OpenAI’s Sam Altman and Elon Musk, publicly called for the pace of development of the most advanced models to slow, citing safety concerns after a run of security incidents involving AI systems and a high-profile resignation from within the industry. OpenAI’s Altman also said the company would not list on the share market in 2026 as had been widely expected.

The initial reaction was sharp. SoftBank, a major investor in OpenAI, fell almost 11 per cent in Tokyo, Korean chipmaker SK Hynix dropped 6.4 per cent and Dutch chip equipment maker ASML lost 6 per cent. The ASX was not spared either, with the sell-off flowing through to local stocks linked to the AI build-out, including parts of the resources sector. US markets recovered most of their early losses by the close, though, and within a week sentiment had swung back the other way. The US semiconductor index rose more than 14 per cent in five trading days, helped along by enthusiasm for a new AI agent launched by Meta, and the Nasdaq pushed to a fresh record.

That round trip says a lot about where the AI trade sits. Sentiment is swinging hard on headlines, a small number of companies are carrying a large share of global index returns, and profitability at the leading model developers is still some way off. OpenAI, for example, reported operating losses last year that exceeded its revenue by a wide margin, according to financial documents reported by the Financial Times. None of that means AI is not real. It clearly is. But it does mean the gap between the long-term story and what investors are paying today deserves some scrutiny.

The Australian angle

RBA Governor Michele Bullock spoke about AI at some length in a speech on 22 September. She described it as the “great white hope” for productivity, a view she said is shared by central bankers around the world, but noted there are very few signs yet of it lifting the supply side of the economy. What it is doing, in her words, is adding to demand through the data centre and infrastructure build, at a time when Australia already has more demand than the economy can comfortably meet. That is part of why AI now features in the rate conversation. Bullock also acknowledged the risk that valuations are running ahead of realistic earnings, saying it might be a bubble or it might not, and that a disorderly unwind would have consequences for the financial system.

Treasurer Jim Chalmers took a similar line in the recently released Intergenerational Report, calling AI the biggest economic transformation of our lifetime and noting Australia is adopting it faster than almost any other G20 country. Locally, data centre operator Firmus is reportedly preparing to raise around $7 billion in what would be one of the larger ASX floats in recent years, with trading expected to begin in late October.

The RBA lifts the cash rate to 4.60 per cent, the highest since 2011

On 29 September the RBA’s Monetary Policy Board raised the cash rate by 0.25 per cent to 4.60 per cent. It is the fourth increase of 2026, it ends a two-meeting pause and it takes the cash rate to its highest level since November 2011. The decision was widely expected. All four major banks had forecast it, as had 33 of the 34 economists surveyed by Reuters, and futures pricing had put the chance of a hike at around 90 per cent. In its statement, the Board said the three increases since the start of the year had tightened financial conditions and the economy appeared to be slowing, but that inflation was still too high and further tightening was warranted to return it to target in a reasonable period. It also made clear it would continue to do what it considers necessary, which leaves the door open to further increases.

The reasoning was fairly straightforward. Trimmed mean inflation, the RBA’s preferred measure of underlying price pressure, was 3.6 per cent in the year to July, unchanged from June and still above the 2 to 3 per cent target band. Headline inflation was 3.5 per cent. Speaking in Sydney the week before the meeting, Governor Bullock said the jobs market was probably still a little too tight, that an unemployment rate somewhere between 4.5 and 5 per cent would take enough heat out of it, and that the board’s priority is stopping the energy price shock from feeding into a second round of domestic price rises. August’s unemployment figure of 4.6 per cent sat at the low end of that range and came with rising participation, so it did little to change the picture.

It is also worth remembering the RBA is not acting in isolation. With the Fed, the ECB and the Bank of Japan all raising rates this month, and global bond yields at their highest levels in close to two decades, there is very little in the international picture pulling Australian rates the other way.

Beyond September

This is where forecasts start to diverge. Ahead of the decision, most economists saw 4.60 per cent as the peak for this cycle, with 26 of 31 in the Reuters poll expecting the cash rate to still be at that level at the end of December. ANZ is the exception among the major banks, forecasting a further increase in November that would take the cash rate to 4.85 per cent, the highest level since late 2008. NAB sees a meaningful risk of a November move without making it their base case, and CBA has said a strong September quarter underlying inflation result could bring one. Futures pricing ahead of the meeting pointed to a cash rate of around 5 per cent by the middle of next year, and no major forecaster is expecting rate cuts in the near term. The August monthly inflation figures, due tomorrow (30 September), and the September quarter figures that follow will do most of the work in deciding whether November brings another increase or a pause.

The banks did not wait for the RBA. Fixed mortgage rates rose sharply in the fortnight before the decision, with CBA lifting its two-year fixed rate by 0.48 per cent to 6.82 per cent. The lowest one and two-year fixed rates from the big four sat at 6.49 per cent in the lead-up to the meeting, noticeably above their lowest variable rates of around 5.99 to 6.25 per cent at the time. When fixed rates are priced above variable, it is a clear signal that lenders expect variable rates to keep rising.

The US midterms: five weeks out, and what history does and does not tell us

Americans head to the polls on 3 November, with every seat in the House of Representatives and 35 of the 100 Senate seats being contested. Republicans currently control both chambers, holding the Senate 53 to 47. Polling has moved toward the Democrats through September, with the national generic ballot average now showing a lead of around 8 points, and affordability and the cost of living are shaping up as the defining issues.

History leans the same way. Since 1914, the president’s party has lost House seats in 89 per cent of midterm elections, by an average of around 30 seats. Polls five weeks out are not results, and plenty can change, but a split Congress, with the House and Senate controlled by different parties or the White House facing an opposing majority, is a realistic outcome.

What it could mean for markets

Less than the headlines will suggest. Morgan Stanley’s research makes the point that the executive branch drives most of the policy that matters to investors, and that the big themes of the past two years, tariffs, geopolitics and deregulation, are likely to continue regardless of who controls Congress. Their analysis also found that, under a Republican president, the strongest average S&P 500 return in the year after a midterm came when Congress was split, at 23 per cent. The logic is simple enough: gridlock tends to mean fewer sweeping policy changes, and markets generally prefer predictability. If Democrats do gain ground, some large technology and energy companies could face more political scrutiny.

Midterm years also tend to be bumpy before the vote. According to Fidelity, the S&P 500 has seen an average peak-to-trough fall of 19 per cent in midterm years since 1961. This year’s largest decline so far, in March, was around 9 per cent. History is a useful guide to the range of outcomes, but it is not a forecast, and this year’s backdrop of an active conflict, high oil prices and rising interest rates is anything but typical.