FED WATCH

My takes on Kevin Warsh’s first FOMC:

• He succeeded in showing (1) a united FOMC (12-0 vote) and (2) his own independence from Trump.
• He was clearly hawkish with much more emphasis on inflation than on employment.
• “We missed the inflation target for 5 years. This committee will deliver.”
• Inflation forecasts were raised for 2026 and 2027. Core PCE went from 2.7% to 3.3% for 2026 and from 2.2% to 2.5% for 2027, all above current consensus of 3.1% and 2.3%.
• Monetary policy is currently “uneven”: “clearly restrictive on housing, tough to say that for anywhere else”. Housing is 16% of US GDP, anywhere else is 84%. Uneven?
• Participants were evenly split between no hikes and one or more hikes (6 for multiple hikes).
• Warsh, not a dot plot fan, did not vote even though he could have kept the 2026 dot from rising.
• Nine members now project hikes by June 2027 vs zero in March. Just one member now sees a cut by June, vs 12 in March.
• Inflation should be measured “to the left of the decimal”. So 2.0% is not a hard number and 2.5% may be acceptable. So “this committee will deliver price stability” is not tied to 2.0%.
• AI may be inflationary over the near term but should drive productivity longer term.
• Fed communications are going back to near-Greenspan style
• Shorter commentaries and explanations. Make-your-own-guess of Fed thinking means more volatility.
• In his view, financial markets should not focus on Fed’s intentions but on actual data. Markets “are the best source of information to judge what policy should be”. And the Fed should use financial markets as inputs. All that sounds a little circumvoluted.
• The FOMC needs better, more up-to-date data.

A good case in point is the all-important employment data. The BLS monthly survey is out the first Friday of each month about three weeks after the reference week that includes the 12th of the prior month.

• The payroll (establishment) survey’s first estimate is preliminary and is based on an incomplete set of employer responses; historically BLS cites initial collection rates around two‑thirds of the sample by the first publication, rising close to 90–93% in subsequent months.

• BLS then issues two scheduled monthly revisions: “second preliminary” in the following month and “final sample‑based” two months after the initial release.

• After those three “vintages,” a larger, once‑a‑year benchmark revision aligns the survey with near‑census administrative data (state unemployment insurance tax records), implemented with the release of January data.

The initial release is the market‑moving one because of its timeliness, but subsequent revisions can be significant as everybody knows.

High‑frequency series from Indeed or LinkUp postings are “observed” web‑scrape data rather than a sampled survey. The platforms continuously update the underlying postings in real time.

This chart plots the JOLTS data (last data point April) with Indeed Job Postings (last data point June 12). The recent trend is rather weak:

The second chart plots Indeed Job Postings against the the YoY % change in the BLS payrolls (last data point May). Yearly payrolls growth has stalled with a weak trend.

Based on Linkup data (through June 14),  US labor demand is in a clear downtrend, -3.9% from its mid-April peak. It has been free falling since mid-May, right when the last BLS survey was done.

Maybe Warsh et al. should be a bit more worried about the labor market.

College graduates must be worried as Apollo Management illustrates below. The unemployment rate for recent graduates is well above pre-pandemic levels and historically very high (my red line), not a sign of a healthy labor market, is it?

OIL PRICES WATCH

Oil prices dropped sharply after the US-Iran MOU (or MOM, memo of misunderstanding), “fueling” hopes for slower inflation. We’ll see how fast tankers can exit Hormuz and reach their final destinations. (Charts from oilprice.com).

We also need to see how OPEC manages production vs netbacks in what could be a completely new oil era.

US production has increased during the war …

… but for exports mainly. Commercial crude oil inventories are at their 5 year low, as are US strategic reserves.

Meanwhile, gasoline stocks are critically low as the driving season begins.

BTW:

• The White House has allowed its waiver of Russian oil sanctions to expire on June 17 after President Trump suggested the US-Iran ceasefire agreement will boost global crude supply, ratcheting pressure on Moscow as its oil exports soared to 6 million b/d in May.
• Another long-term impact from the war: Saudi national oil firm Saudi Aramco is actively seeking to develop more international crude oil storage facilities on the heels of the ongoing US-Iran conflict, currently operating several tank farms in South Korea, Japan and Egypt.

Xi Jinping wants China to boost demand. Why isn’t it working?

(…) “What we’re seeing here is a renewed slowdown in the economy,” said Frederic Neumann, chief Asia economist at HSBC. “Some of the optimism at the beginning of the year is beginning to fade again quite quickly.” (…)

Analysts said the decline in retail sales was also partly being driven by the waning use of a goods trade-in programme, which has heavily affected sectors such as household appliances and autos. (…)

Investment, meanwhile, is tumbling. In addition to the property slowdown, high input costs linked to the war in Iran are weighing on infrastructure investment, according to Logan Wright, an analyst at Rhodium. (…)

The metric also declined last year, after Xi called for a crackdown on excessive industrial competition, and later hit out at wasteful spending. Some analysts accordingly saw that decline as a reflection of under-reporting, as well as possible data revisions, pointing to contradictions with other indicators. (…)

There are few signs of a more substantial policy shift, especially in terms of consumer stimulus. Beijing’s GDP growth target is the lowest in decades but still aims to hit at least 4.5 per cent. (…)

Shanghai Securities News informs us that China is accelerating its nationwide urban renewal efforts through unprecedented financial and policy support from both central and local governments, a push that is expected to unlock trillions of yuan in new investments over the next five years. 

Beijing is also spearheading a major push to spur AI consumption. Another plan (Who got a plan?):

• China will ramp up fiscal support to spur AI consumption while advancing broader measures including smart product rollouts, e-commerce integration and service upgrades. 
• Authorities will coordinate existing funding channels to expand AI adoption in consumer markets, implement policies to promote purchases of digital and smart products. 
• Encourage local subsidies under trade-in programs for next-generation intelligent devices. 
• Provide consumer loan interest subsidies to support AI‑related purchases, expand tailored financial products and services for AI consumption, and leverage a national AI industry investment fund to drive adoption. 
• Plans to promote rollout of next-generation AI devices such as smartphones, computers, smart home systems and wearable technologies, alongside humanoid and service robots for households and elderly care. 
• Deepens AI integration in e-commerce, logistics and retail, supporting smart stores, digital human livestreaming, intelligent customer services and automated delivery pilots. 
• Calls for expanding AI applications across services including tourism, education and healthcare, while improving infrastructure, standards and safety governance.

From the FT:

Goldman Sachs analysts last month predicted that use of AI agents would result in a 24-fold increase in token consumption by 2030 and that the huge rise in demand would exacerbate a shortage of chips over the next 12 to 18 months.

Since the start of the year, Chinese AI models have overtaken their US counterparts in token consumption, according to data from OpenRouter, an aggregation platform that allows users to access multiple AI models.

Meanwhile, some companies have told workers to use open-source models that can be run locally on their own servers or personal devices, reducing the bill they pay to AI labs and cloud providers.

On AI, both the US and China have shifted from regulation towards direct participation. Governments have long subsidised or protected strategic industries.

But what is new is their growing willingness to become shareholders in frontier AI laboratories — treating them less as software start-ups and more as critical national infrastructure.

That distinction carries profound implications. Ownership creates influence. Once governments become equity holders, these AI companies face complex questions that traditional governance frameworks were never designed to resolve: how to balance national security objectives with shareholder returns; how boards should navigate tensions between commercial growth and strategic priorities; and whether they can simultaneously serve global markets while advancing national goals.

Most observers still view OpenAI and DeepSeek primarily as competitors. They are. But they are also expressions of the same structural trend: the world’s two largest economies showing a willingness to move state capital into increasingly strategic layers of the AI stack. The first phase focused on chips, the second on models. The next may concern the ownership and governance of intelligence itself.

From Axios last Friday:

Like its AI policy, the Trump administration’s AI team is taking shape on the fly.

Departures among key White House officials, combined with rapid advances in technology, are shaking up who’s taking the lead on AI policy in the administration.

Silicon Valley figures David Sacks and Sriram Krishnan have served as key architects of the administration’s AI agenda. But with Sacks stepping back from day-to-day involvement and Krishnan preparing to leave, influence is shifting inside the White House to a broader group of officials and aides.

Here’s who’s running the show, — for now.

Commerce Secretary Howard Lutnick. Lutnick’s signature was on the letter that sparked the latest confrontation between Anthropic and the administration, ultimately leading to the takedown of the company’s Fable and Mythos models. Last week, Lutnick imposed export controls on Anthropic, effectively creating a licensing regime that could eventually impact other AI labs.

Chris Fall at Commerce’s Center for AI Standards and Innovation has been holding technical meetings in D.C.

Treasury Secretary Scott Bessent. Bessent does not directly oversee AI testing. Yet he was the point of contact when Amazon raised concerns about Anthropic safety issues and was among the few Cabinet members flanking the president during a G7 press conference.

Bessent is viewed as “the more reasonable actor” compared to Defense Secretary Pete Hegseth, who has been public about his disdain for the company, said one source familiar with the administration’s thinking.

“Bessent is trusted by the private sector and critical infrastructure operators as a sober actor,” another source familiar said.

White House chief of staff Susie Wiles. Wiles, a veteran political strategist, has not typically been involved in day-to-day AI policy.

But she was receptive to Bessent’s concerns about how Anthropic’s Mythos could impact the financial sector, helping to reopen lines of communication with the company.

National Economic Council’s Ryan Baasch. Baasch is said to be the person carrying the torch inside the White House for Sacks and Krishnan.

Baasch has worked alongside the two advisers to influence AI policy on Capitol Hill and to push federal preemption of state AI laws.

National Cyber Director Sean Cairncross is also holding technical meetings in Washington, but a clash with Bessent over the administration’s AI response has deepened following the latest Anthropic dispute, sources familiar tell Axios.

Cairncross believes that Treasury has become too involved, while Bessent and allies in the White House believe Cairncross has not met the moment with the necessary urgency.

Cairncross’ head of policy and senior adviser Thomas Lind plans to leave, further depleting the Trump administration of technical expertise.

None of the above are engineers.

From the U. of Virginia’s Data and Policy Institute:

Artificial intelligence is at the core of the Chinese Communist Party’s (CPC) vision for national security and technological self-reliance.

Over the course of decades, Chinese leaders have developed layers of national policy that align government agencies, private companies, the military, and academia on the development and implementation of this emerging technology.

China’s AI ecosystem is rapidly developing technological breakthroughs that expand its global influence—and the political control of the CPC. Today, China’s centrally planned AI strategy has made it a peer competitor with the US, whose less regulated AI ecosystem is mostly led by private tech companies. 

The CPC built and layered its AI strategy atop decades of ideological and institutional planning. From Deng Xiaoping’s reform era to Xi Jinping’s digital governance, China developed AI not as a standalone policy initiative, but rather folded it into evolving doctrines of modernization, centralization, and techno-political control.

While not himself an engineer, Deng Xiaoping since 1978 deliberately promoted technically trained specialists, engineers and technicians, into positions of power, seeing technical expertise as essential for modernization; this shift is what produced the later era of engineer‑leaders like Jiang Zemin (1993-2003), Hu Jintao (2003-2013), and Xi Jinping.

In the US, the only true engineer‑president was Herbert Hoover, a mining engineering from Stanford in … 1895.  Jimmy Carter, though not formally engineer, had substantial engineering training and practice in the US Navy.

Not that engineers are better politicians, but understanding technology is clearly a plus, particularly nowadays.

Goldman Sachs:

The Chinese government is restructuring the economy by shrinking the property sector and bolstering the technology sector. This transformation is clearly reflected in physical output data. Compared with 2019 levels, production of industrial robots has tripled, and semiconductor production has more than doubled, in sharp contrast to the decline in construction materials such as glass and cement.

The divergence is also visible in equity market performance. The Shanghai Composite Index has been largely flat year-to-date, but beneath the stable aggregate index, the information technology sector index has rallied by over 50%, while the consumer discretionary sector index has slid by more than 25%. (…)

According to NBS officials, high-tech manufacturing value-added grew 15.1% yoy, far outpacing overall IP growth of 4.5% yoy in May. In contrast to the 4.1% yoy decline in overall FAI in January-May, investment in the information transmission industry (which includes AI computing), R&D, and high-tech sectors increased 30.4% yoy, 9.3% yoy, and 4.5% yoy, respectively.

On the surface, China’s AI capex appears to fall far short of that in the US. (…). However, this does not mean China’s computing power is only a fraction of that in the US. As of mid-2025, installed data center capacity in China was already 60% of that in the US.

Two factors contributed to China’s low hyperscaler capex relative to its data center capacity. First, the cost of building data centers is much lower in China than in the US. Second, a significant share of China’s AI-related capex is conducted by the government. In the 15th Five-Year Plan (2026-30), for instance, policymakers plan to invest RMB2tn in computing power networks, including data centers. China’s push for AI development and adoption is real, even though the way it is financed and applied differs significantly from the US.

Turbocharged Earnings Are Pushing Stocks Higher. There’s a Catch. The AI-investment boom is leading to big spending, but the bill comes later

The stock market is salivating over big increases in corporate earnings, but there’s a catch. Much of the growth is due to a lag in when the costs of huge AI investments hit the books. (…)

When Nvidia sells its chips, for instance, it books the revenue and earnings quickly. But its end customers treat the purchases as capital assets, meaning they defer the upfront costs and recognize them only gradually as depreciation on their income statements.

So, while the outlays immediately hurt the customers’ free cash flow, they get to spread out the acquisition costs over many years on their income statements. In addition, they might not have to start recognizing depreciation expenses until long after the purchases if, for instance, the equipment is going into a new facility still under construction and isn’t being put to use immediately.

Todd Castagno, an accounting analyst at Morgan Stanley, calls the current period “a golden window where everybody looks good.” Revenues and margins look strong across the AI ecosystem at both the hardware suppliers and the big-spending buyers.

There is nothing unique here about the accounting treatment. This is the normal way assets get booked and then gradually written down. What stands out is the sheer size of the capex boom.

A big wave of depreciation expense is coming, and the size of the hit to hyperscalers’ income statements is far from clear. Estimates of hyperscalers’ future capital expenditures keep climbing sharply, and analysts’ estimates of future deprecation expenses diverge widely. (…)

Companies have wide discretion to shorten or lengthen their fixed assets’ useful lives, which changes their yearly depreciation numbers. And a lot of the data-center build-out is being financed off-balance-sheet, which adds to the complexity. Many analysts don’t even attempt to isolate the specific depreciation of AI equipment, and instead lump it into a broad D&A estimate. (…)

The consensus view for now is that capex growth will taper off after next year while revenue keeps surging, allowing free cash flow to rebound in a V-shaped recovery. But nobody really knows how this will play out.

The stakes are high for investors trying to make sense of the numbers, as well as for passive index investors. The forward price-earnings ratio for the S&P 500 using estimates for the next 12 months is about 22 times earnings, which is above historical averages even before depreciation expenses ramp up.

A lot of today’s earnings are coming from spending that won’t appear on the hyperscalers’ income statements for years. Everyday investors have a great deal riding on whether the AI titans can find the revenue someday to justify those costs.

Paul Kedrosky

• Strip Out AI and Energy, and the S&P 500 Is Down

Since January, the entire S&P 500’s gains have come from just two corners of the market, AI and energy, while everything else is actually trading at less than where it started, see chart below. (Apollo)

Retail Investors Remain the Market’s Strongest Source of Demand

Through our unique position as the #1 US Retail Market Maker [Citadel Securities], executing ~35% of all US-listed retail volume, we continue to observe the highest and most persistent levels of retail participation on record. At the same time, the retail investor is evolving. Unlike previous periods of retail enthusiasm, today’s retail flows increasingly resemble those of institutional investors.

Retail trading activity has officially entered a new regime. The defining characteristic of today’s retail investor is not just enthusiasm, but persistence.

May shattered previous activity records in cash equities, surpassing the prior monthly high set in January 2021 by more than 10%. Retail cash equity volumes ran 60% above the 2025 average and more than twice the 2024 average. From this peak, activity has accelerated further in June, with volumes this month tracking 9% above May’s record.

Nine of the ten largest retail trading days ever observed on our platform have occurred in just the last month, including seven during the first half of June alone. Friday (June 12) marked the largest single day of retail net buying in our dataset, surpassing the previous record by 50%.

 

Options activity is similarly breaking records. Average daily options volume on the Citadel Securities retail platform reached a record high in May, running 20% above the trailing one-year average. Participation has continued to build in June, with volumes surpassing the highs made last month by another 20%.

New records are being set almost weekly. The first week of June established a new high for retail options activity, only to be surpassed the very next week. This week is on track to challenge those levels once again.

Retail investors are also deploying more capital, trading a record $5.8 billion of options premium per day during the month of May on our platform. This has already climbed to $7 billion in options premium per day in June.

Rather than only chasing speculative corners of the market, retail investors are increasingly concentrated in the same companies driving benchmark returns and institutional positioning.

Semiconductors sit at the center of this trend. Average daily options premium traded in semiconductor names reached a record high of $1.6 billion in May, more than doubling April levels, running nearly 5x the historical average, and surpassing the previous record by more than 25%. June is already tracking at approximately $1.9 billion per day, 16% above May’s record. (…)

Equity ownership continues to broaden across income cohorts, with the fastest growth occurring among households that historically had the lowest market participation rates.

Since 2010, equity and mutual fund ownership among the bottom 50% of households has increased by more than 570%, outpacing every other wealth cohort. Today, the bottom half of households owns more than $600 billion of equities and mutual funds, a record high.

Despite increased engagement, households hold record cash balances, waiting to deploy capital on market dips. This dynamic only changes when the VIX rises above 30. Today, the VIX is approximately 16.

We observe this “buy-the-dip” behavior on most down trading days in our data at Citadel Securities.

 

Meanwhile, the largest buyer of U.S. equities remains active. Year-to-date, US corporates have authorized more than $925 billion of share repurchases, the strongest pace ever recorded through this point in the year. Technology and Financials account for roughly 57% of all announced buybacks in 2026, reinforcing demand in many of the same sectors already benefiting from strong retail participation and passive flows. Equity issuance remains small even when compared to current buyback demand.

BTW:

 

So, how much cash is there, really?

• Cashed Out: another echo on this theme is investor cash allocations bouncing along the bottom (which has often been a topping sign in the past).

 @i3_invest via @TheChartReport

• Institutions are cash neutral:

@neilsethinew

• No cash? No worries:

The rate of expansion in margin debt is the 4th fastest in recent history, and is consistent with a heightened downside risk alert. I would note that the pervious 3 major topping signals on this indicator did take a few months to work, but the message is pretty clear here. (Callum Thomas offers the next 2 charts)

Source:  Topdown Charts Professional

• Similar to what we saw with investor allocations to cash probing the lows, investors in aggregate are running record low allocations to treasuries.

• Feeding the beast …

• … while the beats wants to be fed:

Since 2002, there have been 17 “sell signals” where the average loss for global stocks over 2-3 months is 2-3%, with a hit ratio of ~60% and max drawdowns of 15-20%. – Michael Harnett, BofA

• Also feeding the beast:

Risk appetite likewise remains hearty in more speculative credit categories, with domestic leveraged loan funds attracting net inflows for 10 consecutive weeks as of this past Wednesday. 

As Bloomberg reports, Wall Street banks are capitalizing on that friendly backdrop to clear their books of “a raft of leveraged loans that they previously couldn’t sell,” with at least a trio of so-called hung deals coming to market of late.  Private equity-promoted firms, meanwhile, have sold $2.9 billion in loans this month with which to pay themselves dividends, per a Tuesday bulletin from PitchBook, already marking the busiest monthly period for such recap deals since January.

Though index-level spreads across investment-grade and high-yield credit remain near their tightest of the post-Lehman Brothers epoch, storm clouds gather above the deep end of the speculative pool. Triple-C-rated bonds tracked by Bloomberg settled yesterday at a 748-basis point option-adjusted spread, up more than 100 basis points from the start of May to mark the largest premium since the Liberation Day episode of spring 2025. (ADG)

This beast also needs feeding:

Pentagon Tells Lawmakers It Needs $80 Billion for Iran War and Other Bills Military officials say the services could run out of money for operations this summer unless Congress passes a new wartime spending bill

FYI:

A trust recession is undermining citizens’ confidence in the future & national pride

Developed nations around the world are suffering from a trust recession, nowhere more than the U.S.

Most Americans don’t trust the government to do what’s right, believing the system rigged by self-dealing elites and corrupt institutions. Per NBC News, “The lack of trust in institutions is also taking a toll on American pride.”

Whereas 85% were “extremely or very” proud to be American in 2006, and 81% said the same in 2016, only 56% say so today.

So What? As long as voters perceive a rigged system they’ll keep voting for “wrecking ball” candidates promising radical change. (Bruce Mehlman)

Warsh Overhauls How the Fed Talks and Keeps Markets Guessing on Rates

Kevin Warsh used his first meeting as Federal Reserve chairman to put his stamp on how the central bank operates and communicates.

He trimmed its policy statement, declined to submit a rate forecast of his own, and launched five task forces to study everything from how the Fed communicates to how it analyzes the economy—a sweeping reach into how the central bank operates.

About the things markets most wanted to understand—how the new chairman reads the economy and translates it into policy—he said almost nothing. Pressed later on inflation, on whether policy was restrictive and on the future of the so-called interest-rate dot plot, Warsh repeatedly demurred. “We have a task force for that.”

On one point he was emphatic: The committee was united and determined to bring inflation down. “We’ve missed for five years, and we’re going to fix that,” he said, pledging that the committee would “unambiguously and unanimously” deliver price stability.

But he declined to say whether that meant raising rates. “The good news is we’ll be meeting in six weeks.”

By removing the so-called forward guidance (i.e., verbal or written hints) and sitting out the projections, Warsh put aside the tools his predecessors used to steer expectations. That left the committee’s direction an open question at a moment when its members are split down the middle. Nearly half thought rates should be at the current level by year-end, while the other half thought they should be higher. (…)

Telling markets what the Fed might do at its next meeting is one thing, said Michael Feroli, chief U.S. economist at JPMorgan. Explaining the framework by which it responds to a hotter economy or rising prices is another. “That’s not forward guidance. That’s having a framework,” he said.

Warsh, he argued, had used objections to the former to quietly sidestep the latter. By the end of the news conference, Feroli said, he was no closer to knowing how the committee now reasoned about policy. “It matters to have some sense of how this committee is thinking about how it goes about conducting this business,” he said.

That kind of explanation is standard among major central banks. The Bank of England and the European Central Bank both publish detailed accounts of their reasoning, and Warsh’s three immediate predecessors used the postmeeting news conference largely for that purpose. (…)

The meeting “ushered in a new era of monetary policy” in the U.S., said Rick Rieder, BlackRock’s chief investment officer for fixed income, whom President Trump interviewed this year for the job that went to Warsh. Rieder said less communication could reduce volatility if it builds confidence. (…)

The official statement’s only 2 sentences:

Economic activity is expanding at a solid pace despite elevated uncertainty that owes, in part, to the conflict in the Middle East. Productivity growth and capital investment are strong. Job gains have kept pace with the workforce, and the unemployment rate has changed little.

Inflation remains elevated relative to the Committee’s 2 percent goal, in part reflecting supply shocks that have driven price increases in certain sectors, including energy. The Committee will deliver price stability.

A solid economy, an OK labor market but too high inflation. “The Committee will deliver price stability.”

Then in his presser, to make sure the message was clear: “The commitment to deliver is strong, unanimous and unambiguous,” he said in response to a reporter’s question. “And that’s, I think, an important message. We missed for five years, and we’re going to fix that.”

That’s clear guidance for me.

So was the dot plot:  Nine of 19 officials penciled in at least one rate increase by year’s end, up from none in March. Just one foresaw a cut, down from 12. The average projected rate by the end of this year has risen from 3.24% to 3.83%.

“We have the capability and commitment to deliver on our price stability objective,” said Warsh. “That’s exactly what we’re going to do.”

Warsh and the FOMC could have used the fall in oil prices to duck the inflation risk. There was none of that. Maybe yesterday was Kevin Warsh’s independence day!

Rob Kaplan, vice chairman at Goldman Sachs Group Inc. and former Dallas Fed president in Bloomberg:

“If inflation prints don’t cool between now and we get to September, I actually think the balance of risks suggests it would be wise to take some action, either in September or in the fall,” Kaplan said in a Bloomberg TV interview. “That’d be the wiser thing to do.”

He also noted that Fed policy moves rarely occur as one-off actions, with rate changes more often coming in a series of two or three. “So I think if you move in September, you need to be prepared. There could be one or two more,” he added.

Kaplan cautioned against reading too much into the Fed’s latest dot plot, saying it might not have reflected the US-Iran deal and reopening of shipping routes. The outlook could look different by the time officials release their next set of forecasts in September.

“If I were in my former seat, I would be urging caution about interpreting this dot plot because we just had a big change and I want to give a chance for that to work through the system,” he added.

Warsh also said:

I think financial markets perform best when they react to incoming data. I think the financial markets work less efficiently when they ask a question, ‘How will the Federal Reserve react to that incoming information?’. Financial markets can offer valuable data to central banks—but not when investors are betting primarily on what the central bank itself will do.

Also yesterday:

US Retail Sales Rise in Broad Gain Despite Higher Gas Prices

US consumers ramped up spending in May across a broad range of retailers, signaling households powered ahead despite higher gasoline prices.

The value of retail purchases rose 0.9%, marking a fourth straight month of increases, Census Bureau data showed Wednesday. The figures aren’t adjusted for inflation.

Outside of receipts at gas stations, sales rose a firm 0.7%. (…) So-called control-group sales — which feed into the government’s calculation of gross domestic product — were up 0.7%. That measure excludes food services, auto dealers, building materials stores and gas stations.

“Spending is pretty broad-based. This isn’t a fluke,” said Shruti Mishra, a US economist at Bank of America Corp. “The consumer appears to have made it through the worst of the energy shock relatively unharmed.” (…)

The figures are consistent with card data from Bank of America, JPMorgan Chase & Co. and PNC Financial Services Group, which indicated spending continued to advance last month alongside larger-than-usual tax refunds and a rising stock market. (…)

Spending at restaurants and bars, the only service-sector category in the retail report, fell 0.1% in May after a strong increase in April.

This chart shows how nominal retail sales growth shot through growth in labor income in recent months as my retail inflation proxy jumped from 1.0% YoY to 5.0% in just a few months.

How broad?

The Census Bureau lists 11 retail categories other than motor vehicles and gasoline. The five categories that had growth below the 5.6% average for May grew only 2.0% YoY on average. On a MoM basis, 7 groups had growth below the +0.5% average for a 0.06% MoM growth rate, essentially flat in May.

Street economists may not feel it but Main Street Americans must deal with food inflation of 3.0-3.5% (food-at-home was +4.2% a.r. in April-May) or apparel up 4.8% (+5.5% a.r.) including footwear at +5.2% (+12.0% a.r.).

Inflation on furniture/home furnishings and building materials is tame but only because demand is so weak.

Trump Says Anthropic Negotiations Continue as AI Leaders Huddle at G-7

(…) At the lunch, OpenAI Chief Executive Sam Altman and Demis Hassabis, the head of Google’s DeepMind AI lab, were seated next to Trump. Amodei was seated on the opposite side of the table from the president, next to French President Emmanuel Macron. Down the table were Lutnick and Treasury Secretary Scott Bessent, who were among the officials that targeted the company’s models last week.

The ban on Anthropic models raised concerns among industry executives and foreign governments that unilateral actions by the administration could dictate who can use leading AI models.

Macron told reporters the U.S. made a strong decision that reflects growing awareness that AI can be dangerous. “The bad thing is that the reaction is, in a way, strictly nationalist,” he said, adding that democracies need to cooperate on these issues.

“We won’t buy any models made by these companies if overnight, you can just flip the switch,” Macron said. (…)

At the lunch, Amodei and Hassabis advocated for the U.S. to lead G-7 countries in shaping AI safety standards, people familiar with the meeting said. Others proposed a more decentralized standards-alignment process, similar to that of the Financial Stability Board set up in the wake of the 2008 financial crisis.

Altman called for an international forum to settle on “globally accepted standards for testing” and “impartial analysis” of AI risks, adding that “it is crucial that we do not allow the risks of this technology to lead to undue concentration of power,” according to excerpts of his remarks.

“One thing everyone agreed on is that there does need to be an effort initiated to coordinate among the G-7,” said Aidan Gomez, chief executive of Canadian AI startup Cohere. (…)

One European leader said that the case is becoming compelling for good-enough homegrown tech that costs less than the US version and has fewer strings attached. China must be seeing an opening with its low-low cost open source models.

The FT quotes Arthur Mensch, CEO of French AI company Mistral: “one of the big concerns discussed was how countries could make sure they had a supply chain for AI technology they could control. When you have an intertwined supply chain, are you sure that your counterparts can’t cut you off? That was mentioned multiple times mainly by non-US participants”.

It’s not only LLMs and other software apps, it’s the whole supply chain that’s seen at risk, even among “allies”.

Trump Demanded Iran’s ‘Unconditional Surrender.’ He Got a Surprise Instead.

(…) But it was Mr. Trump himself who offered what may be the most cleareyed answer about why he needed to end this war so fast. He didn’t want comparisons to Herbert Hoover, he told reporters at the Hotel Royal in Évian-les-Bains, on the shores of Lake Geneva, on Wednesday.

“He was always the one I didn’t want to be,” Mr. Trump said of the 31st president, who presided over the market crash that ushered in the Great Depression. “I didn’t want to see economic catastrophe.” Later he noted that if the war continued, the world would have begun to run out of oil stockpiles.

That combination — economic chaos and disrupted oil markets — is exactly what the Iranians viewed from the opening days of the war as their most potent weapon. They executed on that vision with precision, closing the strait and blowing up petrochemical facilities, desalination plants, hotels and air bases across the Gulf. And by the president’s own testimony, it worked.

If that was Phase 1 of Iran’s strategy, history suggests Phase 2 may be one of delay and more delay. In past negotiations, the Iranians refined the art of arguing over every paragraph, throwing in new obstacles to inspections or reinterpreting the meaning of “nuclear research” to embrace continued uranium enrichment. Few were more skilled at this process, former American negotiators say, than Abbas Araghchi, the Iranian foreign minister, and a veteran of past talks.

And Mr. Trump, eager to move on, seems to be paving the way for a long, slow process. On Tuesday, he said he wasn’t especially concerned with getting Iran’s nuclear fuel — now buried under the rubble of last year’s American air attacks — out of the country. On Wednesday, he acknowledged the talks would probably go beyond 60 days. (…)

“The only ‘achievement’ of the ceasefire is the likely re-opening the Strait of Hormuz — which was open before the war started,” former Secretary of State Antony J. Blinken wrote online on Wednesday. “And we will apparently pay Iran to do so, in the form of waivers for the export of Iranian crude oil. Iran has now demonstrated the capacity to stop or slow the passage of oil, natural gas, fertilizer and other critical products upon which so much of the world depend.”

Mr. Blinken, an architect of the 2015 accord, concluded: “Going forward, it will almost certainly find ways to collect ‘fees’ for safe passage that will help entrench the regime.” (…)

But the bigger risk may be this one: When Iran’s leaders begin to clear the rubble left by 40 days of bombing, and think about how to spend the billions in oil revenue that will soon resume, they may well question whether they had the right nuclear strategy. (…)

On Sunday, when Mr. Trump called The Times, this reporter asked him whether Iran might now follow the North Korean model. “He’s got serious nuclear weapons,” Mr. Trump said of Kim Jong-un, whom he threatened with annihilation during the first Trump term, then met three times in a fruitless effort to convince him to disarm. “But that should not have been allowed,” he said, asking whether North Korea got the bomb under President Clinton or President Obama. (It made its first test under President George W. Bush.)

But Mr. Trump evaded the question of whether his decision to attack Iran could ultimately drive it to follow North Korea’s model. And he insisted his deal would stop Iran, saying Prime Minister Benjamin Netanyahu should thank him for keeping Israel from nuclear annihilation.

“Whatever it takes,” he said. “Forty-seven years,” he said, referring to the 1979 Iranian revolution, “nobody was able to do it. And we did it. We did it the right way.”

History may prove him right, but it is far too premature to make that claim. Maybe even he knows that, based on his statements on Wednesday morning. If the accord didn’t stick, he had a plan, he insisted. He would “go back to bombing.”

Like if that plan worked…