The third quarter opened with the Federal Reserve’s first expected rate cut sliding further into 2027 and an economy warming at the top line while the workers inside it felt none of it. It closed with a unanimous quarter-point increase, crude near $110 a barrel and consumer confidence at its lowest level since April 2014. In 90 days the governing question moved from when the easing would begin to how many increases the year still holds.

The deeper story was about a central bank that stopped looking through supply shocks. Three of them overlapped through the quarter: an energy crunch that reopened in July and worsened in September, an artificial intelligence (AI) buildout that lifts investment and imports in nearly equal measure and the residue of tariffs. For most of the past two years the committee treated disturbances of this kind as noise to see past. By the end of September, it had decided otherwise, and the decisive input was not a wage number or an output gap but a damaged pipeline in Saudi Arabia. The quarter also closed with a set of annual revisions that made the recent past look better in almost every respect and changed the path ahead not at all.

Executive Summary

• Rates are on the rise. The Federal Open Market Committee (FOMC) held the policy rate at 3.5% to 3.75% in July on a 9-to-3 vote, with three regional Federal Reserve presidents dissenting in favor of an increase, then raised the rate 25 basis points to a range of 3.75% to 4% in September, unanimously, with the projections pointing to risk management rather than overheating as the motive.
  
• Energy struck twice. The collapse of the United States-Iran truce reclosed the Strait of Hormuz in July and pointed retail gasoline back toward $4.50 a gallon, and September damage to Saudi Arabia’s East-West pipeline, the critical alternate route for Middle East crude, sent oil close to $110 a barrel. Diesel became the most consequential price in the economy at $6.50 a gallon, a level consistent with crude at $156, with tightness in refined products capable of adding as much as four tenths of a point to core inflation next year.
  
• The labor market maintained its equilibrium. The labor market stopped hiring without starting to fire. Nonfarm payrolls fell 23,000 in July before rising 162,000 in August, while the layoff rate slipped to 1.0% and the hiring rate held near 3.3%.
  
• Annual inflation revisions rewrote the recent past favorably. Core Personal Consumption Expenditures (PCE) inflation now runs at 3% rather than the pre-revision 3.3%, growth in 2024 and 2025 came in two tenths higher with most of the gain in AI-exposed categories, and the personal saving rate moved up by more than 1.5 points to above 4%.
  
• The consumer maintained its split personality. Households spent well and felt terrible. Confidence fell to 81.9, its weakest since April 2014, sentiment finished the quarter at 48.1 with year-ahead inflation expectations at 4.6% and the third quarter nowcast for consumption still ran at 3.3% annualized.
  
• AI giveth and taketh away. The AI buildout kept adding to activity and subtracting from measured output, with capital goods imports up 57.2% over the year and net trade on track for a 3-percentage point drag on third-quarter growth, the largest since the first quarter of 2025.

The Fed: From Waiting to Hiking
No institution traveled further this quarter than the Federal Reserve, and it traveled in the direction nobody had penciled in. July opened with the first expected cut sliding toward the back half of 2027, pushed there not by a cracking economy but by a steady one that gave the committee no reason to hurry. At the July 29 meeting the committee held at 3.5% to 3.75%, and the vote told the more interesting story: 9 to 3, with three regional Federal Reserve presidents dissenting in favor of an increase, against market pricing of roughly a one-in-three chance of a hike, which sent Treasury yields sharply lower once the decision landed. A committee that argues loudly and then leaves the thermostat alone is still a committee that has moved, if only in temperament.

Jackson Hole was the hinge. Chair Kevin Warsh used his keynote to reaffirm 2% inflation on the PCE measure as a firm, fixed target. After months of studied ambiguity, the implication from the Chair is that policy settings were not restrictive, a departure from the language of his predecessor and from a dot plot placing the policy rate above the longer-run neutral level, and that underlying inflation was moderating too slowly. Markets did the rest, moving the September decision to a coin toss from roughly one-in-three odds and pricing two increases over the coming year. A chair who declines to give forward guidance had moved the rate path further with one speech than most predecessors had managed with a statement.

September delivered what August had implied. The committee raised the policy rate a quarter point to a range of 3.75% to 4%, unanimously, with the projections suggesting risk management rather than an overheating economy as the motive and with several participants lifting their estimate of the neutral rate, which amounts to conceding that policy had never been as tight as they supposed. Warsh defended the decision by naming three developments, and one of them was not an economic statistic at all, since alongside higher energy costs and a labor market that had stopped worrying him he cited the geopolitical flare-up in the Middle East.

The structural revision matters more than the next meeting. The path now runs to a cutting cycle spread across 2028 that arrives at a new and higher neutral rate of 3.25% to 3.5% A change that reflects stronger productivity growth, AI-driven investment demand and persistently large federal deficits that is likely to keep longer-term Treasury yields elevated on its own. Markets have run further still, pricing 90 basis points of tightening over the coming year, which Oxford Economics doubts the committee will deliver. The quarter therefore ends with the same disagreement it spent September resolving, only shifted one increase to the right.

• Key Takeaway: The Fed spent the first half of the quarter finding reasons to wait and the second half finding reasons to move, and the trigger came from outside the domestic economy entirely. What changed durably is the level around which policy settles, since a higher neutral rate is likely to keep long yields elevated regardless of how many increases December brings.

Energy: Two Shocks in One Quarter
The energy story that Q2 closed on a note of relief reopened within days of the new quarter. The truce between the United States and Iran fell apart in early July, the Strait of Hormuz closed again and wholesale prices pointed retail gasoline back toward $4.50 a gallon after weeks of improvement. By the end of July gasoline had climbed back above $4, keeping upside risk to inflation alive at precisely the moment the committee would have preferred a quiet summer.

September brought the larger blow, and it arrived from infrastructure rather than diplomacy. Damage to the East-West pipeline in Saudi Arabia, which had served as the critical alternate path for Middle East oil exports since the Strait closed, pushed crude close to $110 a barrel and looked to keep prices higher for longer. This had economists reworking their baseline forecasts, lifting West Texas Intermediate forecasts to an average above $100 a barrel in the fourth quarter against $80 previously. A single length of pipe moved the oil forecast by more than twenty dollars a barrel, which is a useful reminder of how little of the macro outlook the macro data actually contain.

The most dangerous price was not crude but diesel. It reached $6.50 a gallon, a level consistent with an oil price of $156 a barrel, a gap that speaks to an extraordinary refining spread and that will persist while tanker traffic through the Strait runs far below normal. Diesel matters more than gasoline for the inflation outlook because it sits inside the cost of producing and shipping ordinary goods and feeds directly into grocery, restaurant and delivery prices. Tightness in refined products could add as much as four tenths of a percentage point to core inflation next year. The proposed remedy carries its own hazard, since a ban on diesel exports would lower prices on the Gulf Coast and in the Midwest while raising them in the Northeast and on the West Coast, and would risk cutting refinery output and lifting gasoline prices. Hopefully, the measure is unlikely to occur.

The forecast arithmetic followed directly. The revised baseline carries headline Consumer Price Index (CPI) inflation as much as four tenths higher and nudges core CPI up a tenth, with food inflation now peaking later in 2027 and the additional energy passthrough to core prices running mainly through non-energy consumer goods and airfares. Firmer prices arrive paired with slightly slower growth, at 2.1% this year and 2.4% next, each a tenth below the estimate at the start of September.

• Key Takeaway: The quarter contained two separate energy shocks, one diplomatic and one physical, and the second proved worse than the first. Diesel is the variable that carries the shock into everything else, and it is the reason a supply disruption half a world away now shows up in a domestic policy rate.

Inflation: The Round Trip That Did Not Finish
Inflation spent the quarter looking as though it had turned the corner, right up until it did not. June brought a benign reading, with headline CPI falling 0.4% and the annual rate dropping to 3.5% from 4.2% while core prices came in flat, showing little of the tariff or AI passthrough the committee had watched for. July was nearly as kind, with headline CPI up 0.1% and core up 0.2%, leaving core running at 2.5% over the year. Two months of that sort of data is the stuff of which patience is made.

August ended it. Core CPI surprised to the upside at 0.3% with headline CPI up 0.4%, and the strength sat in non-housing core services, the category the Fed watches most closely because it reflects domestic demand rather than imported costs. Airfares, motor vehicle maintenance and repair, lodging away from home and wireless telephone services all accelerated together, which is the pattern energy costs leave behind when they stop being an energy story and become a services story.

Then the statisticians intervened. The Bureau of Economic Analysis (BEA) annual update of the national accounts revised measured core PCE inflation down by roughly three tenths, leaving it at 3% rather than the pre-revision 3.3%, on methodological improvements to three contested categories: computer software and accessories, portfolio management and legal services. The bulk of the improvement sat in portfolio management and investment advice fees, which rise and fall with stock prices and which Fed officials had already discounted in favor of the market-based PCE index. Getting a better thermometer is genuinely useful. It does not make the room any cooler.

Which is why the revision changed the record without changing the outlook. Core inflation still sits well above target with risks stacked firmly to the upside because of refined petroleum prices, and higher energy prices is projected to keep inflation tracking above 3% through the first quarter of 2027. The quarter thus ends with the curious result that inflation was lower than everyone believed and the Fed tightened anyway, which is what happens when a central bank starts managing the distribution of outcomes rather than the central case.

• Key Takeaway: Two benign months gave way to a firm August and then to revisions that lowered the entire recent history of core inflation. None of that resolved the underlying problem, since the level remains too high and the energy passthrough still ahead of us points in one direction only.

The Labor Market: The Expansion That Forgot to Hire
The labor market spent the quarter delivering the same message in progressively stranger ways. July payrolls fell 23,000 against a forecast gain of 80,000, though a 50,000 drop in state and local government employment accounted for the entire decline while private payrolls still rose 30,000. August then produced 162,000 against a consensus of 55,000, with July’s reported decline flipping to a 21,000 gain and the three-month average climbing to 71,000. Two months, two opposite headlines, one unchanged trend.

The reason those headlines meant so little is that the yardstick kept shrinking. A dwindling pool of available workers dragged the breakeven pace of hiring, the monthly job growth needed to hold the unemployment rate flat, toward 20,000 in the second half of the year, low enough that a negative print carries almost no information. The preliminary annual benchmark revision then trimmed 79,000 from the March level of payrolls, a rounding error beside the 861,000 final revision for 2025, and in nine of the past ten years the final revision has landed more positive than the preliminary estimate. Measuring this labor market against a fixed standard is like weighing yourself on a scale that recalibrates each morning.

The mechanism underneath deserves more attention than it received. This year’s improvement ran almost entirely through falling layoffs rather than rising hiring, with the layoff rate slipping to 1.0% in August from 1.1% while the hiring rate managed only 3.3% from 3.2%, inside the depressed 3.1% to 3.5% band of the past two years. The quits rate held at 1.9%, which is what workers do when they doubt they could find something better. Claims confirmed the calm, touching 187,000 in July, the lowest level since 1969, a figure that predates the microprocessor, and September closed with private payrolls in the Automatic Data Processing (ADP) report up 90,000 while the three-month average held at 57,000, close to breakeven. With layoffs already at historic lows, any further warming has to come from hiring, and hiring is the part that energy costs and tighter credit reach first.

For policy, the wage numbers mattered more than the payroll numbers, and they cooled all quarter. Average hourly earnings growth eased to 3.1% over the year while the Oxford Economics wage tracker fell below 3% for the first time in five years. Slower labor force growth is projected to hold the unemployment rate near 4.2% through 2027, which leaves the labor market as the one part of the inflation debate that never became an argument.

• Key Takeaway: The quarter produced a negative payroll month, a strong one and a downward benchmark revision, and none of the three changed the picture of a market balanced at a very low level of churn. Labor never threatened the inflation outlook, which is precisely why the shocks that did came from outside it.

The Consumer and Housing: Stronger on Paper, Sourer in Practice
Households spent the quarter behaving better than they felt, and the gap between the two widened every month. Second-quarter consumption grew a resilient 3.2% annualized on the advance estimate, trimmed later to 2.9%, July retail sales dropped 0.6% on seasonal distortion around a shifted online shopping event and August recovered with a 1.2% gain led by discretionary categories. By quarter-end, after revisions, the third quarter nowcast for consumption ran at a 3.3% annualized pace against 2.8% previously.

The revisions improved the foundation as well as the flow. The personal saving rate moved up by more than 1.5 points to above 4%, meaning households entered the energy shock with a materially larger cushion than anyone had credited them with, and gross domestic income came in higher on new and much larger estimates of net interest income, which leaves recent spending resting on firmer ground than the pre-revision data implied.

None of which the consumers noticed. The University of Michigan sentiment index finished September at 48.1, with the expectations component down 10% for its largest decline since April 2025, and year-ahead inflation expectations jumped six tenths to 4.6%, above the 3.4% level that prevailed before the war. The Conference Board measure fell to 81.9 from 88.6, the weakest reading since April 2014, when the index was still climbing out of the Great Recession. Its labor differential, the gap between those calling jobs plentiful and those calling them hard to get, dropped 2.5 points to 1.7, the lowest since February 2021, which says less about layoffs than about how it feels to look for work when nobody is hiring.

Housing was the clearest casualty of the rate path. Mortgage rates climbed more than 30 basis points over the final month toward 7%, homebuilder sentiment fell three points to 32 on deteriorating current and expected sales and existing home sales slipped 2% in August to 3.98 million, down 1.2% from a year earlier. Prices offered a last piece of good news that reads like a peak, with the national index up 1.9% over the year in July, the largest annual gain since June 2025, though favorable base effects have now passed and the 2027 price forecast is due for a markdown, since rates above 7% price buyers out even as they keep sellers on the sidelines.

• Key Takeaway: The revisions made the household balance sheet look considerably better, and confidence fell to a twelve-year low anyway. Spending is carried by the households that own assets, while the households that buy gasoline set the tone of the surveys and housing absorbs the full cost of a rate path neither group chose.

The AI Economy: Adding More, Netting Less
The through-line of the quarter, as of the two before it, was artificial intelligence. Core capital goods orders rose 1.6% in August on solid gains across primary metals, machinery and electrical equipment, keeping the forecast for equipment spending to expand 10.7% in 2026, the strongest annual pace in more than a decade. Corporate profit margins sit at all-time highs, business inventories remain lean and the tailwind from AI keeps broadening from information-processing equipment into machinery and metals, which is why higher financing costs have not yet bitten.

The same buildout keeps punishing the trade account. The advance goods deficit widened $13.7 billion to $132.6 billion in August, the largest since the first quarter of 2025, as imports climbed 5.5% against a 1.9% rise in exports, putting net trade on track to subtract 3 percentage points from third-quarter growth. Capital goods imports now stand 57.2% above year-ago levels, and the strength has begun to broaden, with all other imports up 11.8% over the year, their fastest since March 2025. A country that buys the hardware abroad and books the spending at home may continue to see the two largely cancel.

The annual revisions finally gave the buildout some statistical credit. Growth in 2024 and 2025 came in two tenths higher, with most of the revision in business fixed investment and roughly two thirds of that concentrated in categories sensitive to AI, including information technology equipment, research and development and data center investment. The third estimate of second-quarter output rose to 2.2% annualized from 1.5%. Taken with the downward revision to measured inflation, the message is that AI has been adding more to growth and less to prices than the data previously showed, which is close to the opposite of how the debate ran for most of the year.

The dependency is the part to keep in view. The revisions do not change the perception that the economy leans increasingly on AI gains and on the wealth effects that lift higher-income spending, which leaves it sensitive to a sudden reversal of optimism. There is a more hopeful reading available as well, since stronger measured productivity could shift the conversation from what AI does to inflation toward what it undoes, with disinflationary effects that may show up more prominently in 2027. Both readings rest on the same assumption, which is that the capital keeps flowing.

• Key Takeaway: AI drove the investment cycle, the import surge and after revisions, a meaningful share of the growth the economy turned out to have had all along. The concentration cuts both ways, since the same force that has carried three quarters of expansion is the one whose reversal would be hardest to replace.

Final Thoughts
The third quarter of 2026 will be remembered as the quarter the Federal Reserve stopped waiting. It began with the first expected cut sliding toward the back half of 2027 and ended with a unanimous increase and another penciled in for December, and the decisive event was neither a wage spiral nor an overheating consumer but a damaged pipeline in Saudi Arabia that took crude toward $110 and diesel to a level implying $156 crude. Underneath the policy drama the economy performed better than it felt, with the labor market balanced at a very low level of churn, wage growth below 3% and annual revisions that raised past growth, lowered measured core inflation to 3% and lifted the saving rate above 4%. The unfinished business is the energy passthrough still working its way into core prices, a consumer whose confidence sits at a twelve-year low while spending holds and an expansion that depends on an AI capital cycle it could not easily replace. The fourth quarter will test whether a central bank that has learned to react can avoid overcorrecting for a disruption that monetary policy was never able to fix, and whether an economy that has absorbed three shocks in nine months has any absorption left.

Oxford Financial Group, Ltd. is a SEC-registered investment adviser. Registration does not imply a certain level of skill or training. The information provided is for general informational purposes only and should not be considered investment, tax or legal advice. Opinions are those of the author and are subject to change based on market, regulatory or economic conditions. Forward-looking statements are opinions and/or estimates and are not guarantees of future results. Data and opinions are based on sources believed to be reliable, including unaffiliated third parties such as Oxford Economics, but their accuracy cannot be guaranteed. Past performance is not indicative of future results. See important disclosures and disclaimers at https://ofgltd.com/home/disclaimers. OFG-2610-07