Quarter · Market review
90 days: the rate pause gave way to September increases
In the rolling 90 days from 4 July to 1 October, a pause in policy rates gave way to September increases, while technology demand faced higher financing and energy costs. Calendar-third-quarter comparisons are separately labelled 1 July to 30 September; they do not replace this rolling interval.
Period covered: 2026-07-04 – 2026-10-01Information as of 14 min read
The 30 September releases included US spending and revised growth data, preliminary German and French inflation, and China’s factory survey. Long-term US Treasury yields ended the month above their August levels.
The US and euro-area policy rate increases made in September remain in effect. Strong sales can support company profits, but energy and new borrowing costs can absorb part of those profits.
Quarter
What happened
The starting point: interest rates had already risen on 17 June
This rolling 90-day period begins on 4 July. Before it began, on 17 June, the European Central Bank’s previously announced interest rate increase took effect. The rate on banks’ overnight deposits rose from 2.00% to 2.25%. These are the initial conditions before the period under review, not a new event within it.
Also on 17 June, the Federal Reserve left the federal funds target range at 3.50–3.75%. It made no change on 29 July. Europe also chose a pause on 23 July. For a business with debt, this meant that energy costs would not be offset by automatically and rapidly falling interest rates.
The starting backdrop: the 18 June agreement had not yet restored shipping
On 18 June, the US announced the signing of a memorandum with Iran. For oil buyers, implementation mattered more: whether ships could sail safely, insurance would be available and oil would actually reach them.
More reliable deliveries would have reduced importing industries’ need to hold expensive inventories. Hauliers would have found fuel and route planning easier. But the agreement itself was not a statistical record of restored flows. July and August data later required a more cautious assessment of precisely this assumption.
July: the supply recovery was not sustained
The International Energy Agency’s 10 July review described a partial recovery in June flows and ceasefire violations on 7–8 July. Its 12 August report said Persian Gulf oil exports in July, including routes bypassing the strait, averaged 15 million barrels per day. That was 2.1 million, or about 12%, below June.
This comparison describes two monthly averages, not a one-day shipping halt. It showed that June’s hopes had not been enough to deliver a lasting improvement. US military statements on 5 and 8 September about strikes on Iranian tankers were a reminder that delivery security remained unresolved. A particular price move cannot be attributed solely to these statements.
23 July: trade rules and sanctions added to the planning burden
The European Union’s 21st sanctions package against Russia expanded restrictions on energy, financial services and the shadow fleet. For companies, this matters beyond politics: payment, insurance and cargo transport options change. But an announced restriction is not a measured fall in exports of the same size.
On 23 July, the Office of the US Trade Representative announced tariff measures on imports from 60 economies, including the European Union. The action concerned insufficient bans on importing goods made with forced labour. Tariffs and exemptions depended on the product and its origin. US importers had to check whether ordered goods would become more expensive, while their overseas suppliers faced the risk that higher final prices would weaken demand.
Late July and August: technology suppliers earned revenue while customers invested
Technology suppliers’ revenue grew while their customers spent heavily on equipment and infrastructure. On 29 July, Microsoft reported capital expenditure, including finance leases, of 41 billion US dollars for the quarter ended 30 June. Cash purchases of property and equipment were 35.8 billion dollars. These amounts must not be added together: the first includes finance leases, while the second shows asset purchases paid for in cash during the quarter.
The company generated 55.4 billion dollars in operating cash flow. Subtracting cash purchases of property and equipment left 19.6 billion dollars. This explains why a rapid pace of construction is not automatically a bad signal: what matters is how much cash the business generates and how much more it needs from outside.
On 26 August, chip designer NVIDIA reported revenue of 96.221 billion dollars for the quarter ended 26 July, up 106% year on year. Operating profit grew 124%. Some customer spending had therefore already become real earnings for the supplier. For customers to earn a return on their investments, however, their new services need demand and must cover equipment, electricity and financing costs.
August and September: growth continued, but costs did not disappear
Both the US and euro-area economies grew in the second quarter. The third estimate of US gross domestic product, published on 30 September, showed annualised growth of 2.2%, or 0.6% quarter on quarter. The second US estimate released on 26 August had been lower. Euro-area growth was revised to 0.6% quarter on quarter on 7 September. On the same quarterly comparison basis, growth in the two regions matched.
The US added 162,000 jobs in August; the data were released on 4 September. On 9 September, second-quarter revenue growth in selected US service industries was revised to 3.0% from the previous quarter. This revenue measure has not been adjusted for prices. Rising sales and a greater real volume of services delivered are therefore not the same thing.
9–11 September: a higher energy-price forecast was followed by an ECB rate increase
On 9 September, the US Energy Information Administration raised its second-half average Brent price forecast to about 90 dollars per barrel. Model inputs were finalised on 3 September. This is a conditional average for a future period, not the oil price in a particular September session and not an already measured effect of later military action.
On 10 September, the European Central Bank announced another interest rate increase of 0.25 percentage points. The deposit facility rate rose from 2.25% to 2.50% on 16 September. On 11 September, the US Consumer Price Index showed annual inflation of 3.4% in August, the same as in July. The seasonally adjusted monthly change accelerated from 0.1% to 0.4%. These data were known before the Federal Reserve meeting of 15–16 September, which subsequently raised the target range.
? How should we read GDP and price indices?
Gross domestic product measures the value of final goods and services produced in an economy. Real growth removes the effect of prices. US quarterly growth is often reported at an annualised rate, as if the same change continued for four quarters. The Consumer Price Index measures a basket of prices weighted by consumption. Its annual and monthly changes have different comparison bases. The Personal Consumption Expenditures Price Index also includes spending made on consumers’ behalf, uses different weights and accounts for changes in the composition of consumption; these two price indices should not be treated as interchangeable.
Mid-September: June’s hopes were not enough to restore supply
The International Energy Agency’s 11 September forecast projected a larger fall in oil supply than in demand this year. This is a forecast, not the final annual result.
For your investments
Let’s discuss your portfolio
16–24 September: US bond yields rose again after a brief decline
On 16 September the Federal Reserve raised its target range to 3.75–4.00%. An earlier European decision lifting the deposit rate to 2.50% took effect on the same day. The US ten-year yield fell to 4.96% on 21–22 September, but reached 5.18% on 24 September. The fall in US bond yields was therefore short-lived.
The preliminary S&P Global US PMI Composite Output Index, covering manufacturing and services activity, reached 58.4 in the release of 23 September, compared with 56.0 in August. Seasonally adjusted initial unemployment benefit claims of 197,000 for the week ending 19 September, released on 24 September, added to the evidence that the economy was not weakening sharply. These are signals from 23–24 September, but neither a survey nor one week of claims is a final measure of an entire quarter’s growth.
? Purchasing Managers’ Index: what does the threshold mean?
The company survey compares activity with the previous month. Above 50 generally indicates expansion and below 50 contraction. A reading of 58.4 does not mean growth of 58.4%; the preliminary measure may be revised.
September’s decisions extended beyond the US and euro area. On 18 September the Bank of Japan set an overnight interest-rate target of 1.25%, effective from 24 September. On 17 September the Bank of England left Bank Rate at 3.75%. These different paths matter for Japanese bonds and yen-funded investments, as well as British companies’ borrowing costs.
25–28 September: uneven demand and trade relief not yet in force
US August durable-goods orders, reported on 25 September, reached $338.6 billion and were virtually unchanged from revised July. Orders for non-defence capital goods excluding aircraft nevertheless rose 1.6%, while shipments increased 0.6%. These are monthly changes adjusted for seasonality, but not for prices. On the same day, the University of Michigan’s final September consumer sentiment index stood at 48.1, compared with 51.7 in August. Business-equipment demand and households’ assessments were therefore sending different signals.
On 27 September, the White House published US and Chinese product lists proposed for lower import tariffs worth about $30 billion in each direction. The US list of imports from China includes plastic tableware and bedding; China’s list of imports from the US includes beef and dairy products. This publication did not bring tariff reductions into force. On 28 September, China reported that August profit at larger industrial enterprises was 4.2% higher than a year earlier; July’s growth was 11.2%. These are year-on-year comparisons; they do not show whether profit rose or fell from July to August. Companies’ profits would depend not only on lower tariffs, but also on selling more goods at profitable prices.
US spending grew, but household income did not rise after inflation
US household spending supported economic growth in August. On 30 September, the Bureau of Economic Analysis reported that spending increased by 0.6% over the month after removing price changes. However, inflation-adjusted disposable income was unchanged. The increase in spending was therefore not matched by growth in current income. Wages, savings and access to credit will matter for whether households can sustain that pace.
The price figures did not show a fresh decline in annual inflation. In the data revised on 30 September, the Personal Consumption Expenditures Price Index was 3.4% higher in August than a year earlier; excluding food and energy, the increase was 3.0%. July’s figures in the same revised series were identical. Comparing July’s earlier published estimate with August’s new figure would confuse a change in historical estimates with a change in prices.
? Personal Consumption Expenditures Price Index: what does it measure?
The index, known as PCE, measures changes in prices of goods and services bought for consumers, including some spending on their behalf. Excluding food and energy helps separate frequently volatile prices; 3.0% is annual price growth, not a monthly change or growth in the quantity consumed.

The US ten-year yield rose from 4.75% to 5.29% during September
Long-term US borrowing costs increased during September. The ten-year Treasury yield stood at 5.29% on 30 September, compared with 4.75% on 31 August, a rise of 0.54 percentage points. Its increase on 30 September alone was 0.03 percentage points. The whole month’s rise cannot be attributed to yesterday’s consumer-spending release.
? Yields and basis points: how should the change be read?
The US Treasury table gives a standardised annual yield for each maturity. One basis point is 0.01 percentage points, so an increase of 0.54 percentage points equals 54 basis points. This is not an investment return: when yields rise, the price of an existing fixed-rate bond generally falls.

Over the full calendar third quarter, from 30 June to 30 September, the ten-year yield rose from 4.44% to 5.29%. That matters for long-duration bond funds and businesses that need to refinance. September’s policy rate increase forms part of the backdrop, but market yields also reflect inflation expectations, bond supply and the compensation investors require for lending over a longer period.
German energy prices are rising faster, but other prices are not accelerating uniformly
On 30 September, Germany’s statistical office estimated annual September inflation at 3.3%, up from 2.9% in August. Energy inflation accelerated from 10.5% to 14.9%. However, annual price growth excluding food and energy remained at 2.4%, while services inflation slowed from 2.8% to 2.7%. A higher headline rate does not mean that all costs have started rising faster.

France’s preliminary September inflation measured under the common European Union methodology, released the same day, rose from 2.6% to 3.4%. European Central Bank Executive Board member Isabel Schnabel stressed the impact of higher energy prices on other prices and expectations on 30 September. This was a new assessment, not a new interest-rate decision: the 2.50% deposit rate has been in effect since 16 September.
? Harmonised Index of Consumer Prices: why can it differ?
The Harmonised Index of Consumer Prices allows price growth across European Union countries to be compared under a common methodology. Its coverage and weights can differ from the national index, so France’s 3.4% figure here is not its national-index reading.
Chinese manufacturing recovers, but orders and profits are different measures
China’s National Bureau of Statistics reported on 30 September that its manufacturing Purchasing Managers’ Index, or PMI, rose from 49.8 to 50.1 in September. The production index strengthened to 51.7, but the new-orders reading edged down from 50.6 to 50.5. Companies increased production, yet the survey did not show a broader acceleration in orders.
The raw-material purchase-price index meanwhile rose from 56.6 to 60.8. This is an important qualification to the stronger production reading: producing more does not necessarily mean earning more. Chinese industrial businesses need to cover higher costs, while European equipment and materials suppliers need actual new orders, not just higher output at their customers.
US crude stocks rose, but finished-fuel inventories fell
On 30 September, the US Energy Information Administration released inventories for the week ending 25 September. Commercial crude stocks stood at 427.3 million barrels, 2.6% above the comparable week a year earlier. Petrol inventories, however, were 7.4% lower and distillates, including diesel and heating products, were 14.9% lower. All three percentages use the same year-on-year comparison principle.
This distinction matters for transport and industrial companies: crude in storage is not yet fuel delivered to a driver. If refineries process less crude, fuel availability can remain constrained even as crude inventories rise. That could support refiners’ selling prices while increasing hauliers’ fuel bills; actual prices and company profits still need to be assessed separately.

Investment consultation
Do your investments work together?
Review where your money is invested, how much depends on a single market and whether your plan fits your goals. Leave with clearer next steps.
See the consultation process and terms on the registration page.
Quarter
What matters now
A bond buyer and a company refinancing a loan face different opportunities
For a new US bond buyer, a higher yield offers the prospect of more future income. The same change is adverse for the prices of existing long-term fixed-rate bonds. Property businesses, utilities and companies building infrastructure with borrowed funds may face more expensive new financing. Payments on previously agreed fixed-rate loans do not automatically change as a result.
Stronger consumer spending supports revenue at US retailers and service businesses. The risk would increase if household income continued to stagnate and spending had to be cut. In Europe, chemical and transport companies paying more for energy face the opposite pressure: the same sales can leave less profit. October’s labour-market and price releases will help establish which force is gaining strength.
September’s dollar effect for euro investors differed from the full quarter
At the European Central Bank’s reference rates, one euro bought 1.1596 US dollars on 31 August and 1.1355 on 30 September. The euro weakened by 2.08% during September. If an asset’s dollar value had stayed entirely unchanged, its euro value would have risen by about 2.12% from currency conversion alone, before costs and without currency hedging. That is a separate exchange-rate effect, not the return on US shares.
The full-third-quarter currency effect was much smaller: from the 30 June rate of 1.1394 to 30 September, translating an unchanged dollar asset value into euros would have added about 0.34%. September’s result therefore cannot simply be applied to the whole quarter. A stronger dollar helped a euro investor holding unhedged US assets this month, but a later euro recovery could reverse that effect.
Quarter
What comes next
- 2 October: the US September employment report. Changes in employee numbers and wages will help assess whether household income can continue to support spending.
- 7 October: US weekly fuel inventories; 14 October: September consumer prices. The question is whether finished-fuel stocks continue to fall and whether higher energy costs affect a broader range of prices.
- 28–29 October: central bank decisions and US growth data. The US central bank decision is scheduled for 28 October; the ECB decision and the first US third-quarter growth estimate for 29 October. These will allow investors to assess actual quarterly growth and the banks’ responses to the new data.
Important. This is general market analysis, not a personal recommendation to buy or sell. Historical data and company results do not guarantee future returns. Forecasts and conditional scenarios may not materialise.
How is the history preserved and updated?
All five periods in this edition share one information date. Week means the latest 7 days, quarter means 90 days, and year means the trailing 12 months. The next dated edition will have a separate URL. This text keeps its date and assessment; later events are not silently inserted into an older edition. Longer periods assess new evidence alongside earlier developments without repeating every headline.


