Bloomberg Markets

AUGUST 2026 SEPTEMBER 2026

VOL 35 NO 4

The Economics Issue

TRUMP, TARIFFS AND A CAN OF CAMPBELL'S CHICKEN SOUP ...p48 AI UPENDS DATING, CAREERS AND EVERYTHING ELSE IN SOUTH KOREA ...p58 SAASPOCALYPSE? SOFTWARE GIVES WALL STREET A CREEPY FEELING ...p64

---

OVERSEAS DUAL TIME CARDINAL POINTS – SOUTH

Reference 7930V, titanium

Calibre 5110 DT/3

---

VACHERON CONSTANTIN GENÈVE ONE OF NOT MANY.

---

UNRIVALED MARKET TIMING

DeMARK®

You know why. We know when.

MARKET TIMING IS EVERYTHING

DeMARK Service for Bloomberg | DeMARK Re/Search 2.0 | DeMARK Bundle

For decades, DeMARK® has been the definitive name in market timing and analytics. The DeMARK Service for Bloomberg provides access to the full DeMARK Indicator® library, while DeMARK Re/Search 2.0 enables custom, real-time indicator scans. Together, these services deliver the most comprehensive market timing analysis DeMARK has to offer, providing the when to your why.

Discover why the biggest names in trading and investing trust DeMARK. For more information, or to start your 30-day free trial, type DEMA.

www.DeMark.com

---

Contents

Bloomberg Markets

VOLUME 35 ISSUE 4 AUGUST / SEPTEMBER 2026

ON THE COVER

Can Corporation of America in Blandon, Pennsylvania

PHOTOGRAPH BY BENEDICT EVANS

48

Trump and the Tin Can

When tariffs meet an American icon By Shawn Donnan

58

AI Country

A technology reboots South Korea By Heesu Lee, Soo-Hyang Choi, Yoolim Lee and Sohee Kim

64

Software Scare

Wall Street trembles before the SaaSpocalypse By Paula Seligson and Michelle Cheng

70

Stock Shy

Europe finds it tough to get citizens to take risks with their savings By Laura Noonan and Aaron Eglitis

---

Contents

Forward Guidance

9 Data 2.0

Looking for private alternatives to government reports

12 Currency Curse

A weak dollar roils Latin America

18 Small But Mighty

Homegrown stock investors supplant foreigners in Asia

20 Price Paradox

Japan's inflation signals economic revival and sparks consumer pain

22 AI Boot Camps

Colleges cash in on the thirst to learn the technology

FFM

26 J. Rothschild Investing

CEO Maggie Fanari on the discount and private assets

30 Consumer Checkup

Tracking US retail sales and other indicators

32 Auction Stress

Is the US's mounting debt pile affecting bond sales?

35 Gulf Opportunities

Regional issuers are on track to sell more debt than ever

38 Geopolitical Risk

Incorporate new metrics into your analysis

40 Asking Questions

Use these tools to interrogate analyses and data

42 Equities Down Under

Introducing Bloomberg's Australia Domestic index family

44 Simpson's Paradox

Look out for this statistical trap

Ideas

76 Taxing AI

A Yale expert on how—and how not—to adapt the tax code

78 Cheat Sheet

The most important functions you should know about right now

80 The FFM Quiz

Test your market knowledge

A Bloomberg Publication

Before it's here, it's on the Bloomberg Terminal. Learn more at bloomberg.com/company

---

PANERAI

INNOVATION FROM THE DEPTHS

LUMINOR

Built as survival instruments for the military, featuring the distinctive crown protecting bridge, an emblem of Panerai's functional design, our watches are tested to resist shock, pressure, and the elements. At sea and on land, they are reliable tools for modern heroes.

YOUR WATCH FOR ANY ADVENTURE.

---

Guest Editor's Letter

The Economics Issue

This issue is brought to you by the letter "K." The idea of a K-shaped economy was popularized in the wake of the Covid-19 pandemic by economist Peter Atwater to describe charts showing diverging fortunes for the rich and everyone else. It's stayed relevant. The Iran war and its disruption of oil flows has pushed up energy prices, pinching many households. At the same time, artificial intelligence and the race to build data centers is spurring a bonanza for others. Protectionism and ongoing financialization are reshaping business and consumer decisions.

You can see it around the globe. In his deeply reported feature "The Ballad of the American Tin Can," Shawn Donnan looks at the confounding impact of US tariffs by following the supply chain behind the humble cans in your pantry. On the Europe beat, Laura Noonan and Aaron Eglitis examine the European Union's efforts to nudge savers toward stocks to power up the region's growth. And in Japan, Yoshiaki Nohara and Sakura

Murakami show how the shift toward investing is helping some cope with inflation's return while those on fixed incomes fall behind.

At the very heart of the global AI boom in South Korea, reporters Heesu Lee, Soo-Hyang Choi, Yoolim Lee and Sohee Kim tell how the surging demand for memory chips is affecting the economy, society and culture. A debate about how to share the spoils is underway there and globally, including how to tax the new industry, which we explore in a Q&A with Martha Gimbel of the Budget Lab at Yale University. We also look at the boom in AI training and the threat new models pose to existing software companies.

We hope the issue helps you navigate this changing economy, whatever shape it takes next.

Malcolm Scott

International economics editor Bloomberg News

Editor

Pat Regnier

Design Director

Josef Reyes

Deputy Editor

John Hechinger

Investing & FFM Editor

Jon Asmundsson

Graphics Editor

Mark Glassman

Visiting Editor

Kate Davidson

Bloomberg Markets draws on the resources of Bloomberg News, Bloomberg TV, Bloomberg Businessweek, Bloomberg Intelligence, Bloomberg Economics, BloombergNEF and Bloomberg LP.

Editor-in-Chief

John Micklethwait

Deputy Editor-in-Chief

Reto Gregori

Advisory Board

Katherine Bell, Sree Bhaktavatsalam, Katie Boyce, Stephanie Flanders, Heather Harris, Simon Kennedy, David Merritt, Chris Nagi, Pratish Narayanan, Michael Patterson, Kristin Powers, Emma Ross-Thomas

Creative Director

Christopher Nosenzo

Photo Director

Donna Cohen

Managing Editor

Loly Chan

Copy Chief

Stacey Shick

Copy Editors

Mark Leydorf, Nicholas Mullan, Brennen Wysong

Production Manager

Susan Fingerhut

Production Associates

Karen Butcher, Melissa Rodgers

Ad Production/Operations

Debra Foley,

Thomas Gambardella,

Daniel W. Murphy,

Catherine Oneill,

James Outhwaite

Global Chief Revenue Officer

Christine Cook

Chief Revenue Officer

Duncan Chater

Head of US Sales

Hayley Romer

Head of APAC Sales

Sunita Rajan

Head of Europe Sales

David Bradford

Head of Middle East

and Africa Sales

Amit Nayak

For advertising inquiries, go to bloombergmedia.com

comments@bloombergmarkets.com

---

Discreet elegance.

A. LANGE & SÖHNE GLASHÜTTE 1/SA

With its compact case size and elegant design, the SAXONIA ANNUAL CALENDAR provides new accents in the field of calendar watches. Intricate details such as the auxiliary dials, slightly bevelled inward, and delicate baton appliques make up a subtly harmonised composition of light and depth,

gracing the dial with a timeless beauty and assuring optimal legibility of the displays for date, day of week, month and moon phase. The discreet design is counterbalanced by its impressive inner values: the newly developed, lavishly finished 491-part self-winding movement. Never stand still.

Discover our collection online or visit one of our boutiques all over the world.

---

WHERE

RISK MEETS OPPORTUNITY

UNCERTAINTY

Whatever the goal, CME Group provides the tools that global market participants need to manage risk and capture opportunities. With 24-hour access to trading in futures, options, cash and OTC across all major asset classes, you can reach new heights with power and precision.

VISIT CMEGROUP.COM/OPPORTUNITY

STRATEGY

VOLATILITY

POSITION

Robert Sánchez

Premier League Goalkeeper

CME Group

Derivatives are not suitable for all investors and involve the risk of losing more than the amount originally deposited and profit you might have made. This communication is not a recommendation or offer to buy, sell or retain any specific investment or service. Copyright ©2023 CME Group Inc. All rights reserved.

---

Forward Guidance

How to Take Measure of An Uncertain World

By ANNA WONG

ILLUSTRATION BY ARIF QAZI

A FEW MONTHS into the second Trump administration, I found myself in a packed room alongside senior officials of the Bureau of Labor Statistics. The topic, "Ensuring the Quality of US Statistics," would have drawn little more than a collective yawn a decade ago.

Since that conference, though, President Donald Trump fired the BLS commissioner, a third of the bureau's senior leadership departed, and the

longest-ever government shutdown interrupted data collection. Meanwhile, artificial intelligence has given researchers quick access to information that can be used to create new measurements. Against this backdrop, Federal Reserve Chairman Kevin Warsh has put together a task force focused on improving the quality and timeliness of economic data indicators.

The urgency to address this issue

has been building for years. During the worst of the Covid-19 pandemic, when I was working at the Federal Reserve, I was assigned to a stint at the White House Council of Economic Advisers. Official data collections were severely disrupted, so we scrambled to find private-sector alternatives: cellphone tower traffic, payroll scheduling services, credit card transactions. No data were off-limits.

Demand for private data has

9

---

since ramped up, mainly because of three powerful forces: the need for timeliness, collection disruptions and a fast-changing economy. Central banks, governments and traders must make decisions in real time, while statistical agencies can improve accuracy only as more information becomes available.

The pandemic compounded those challenges. Emergency lockdowns and safety protocols forced the BLS to suspend in-person data collection, causing a spike in survey nonresponse rate. Even as lockdowns ended, response rates never fully recovered. Real-time data collection problems became a lasting, post-pandemic structural issue.

In 2021, as the Fed was starting to grapple with elevated inflation, those collection problems contributed to cumulative upward revisions of 1.9 million jobs to the government's data. With accurate information, the Fed would likely have raised interest rates sooner. Conversely, in 2024 and '25, payrolls were revised down by more than a million jobs each year, distorting labor strength and delaying rate cuts.

During the government shutdown in the fall of 2025, officials simply didn't publish an October consumer price index. Did the costs of Trump's tariffs peak that month, as Bloomberg Economics' own data indicate? We'll never know.

In the meantime, prices increasingly

moved online, where they can change daily rather than monthly. Geographic differences narrowed as e-commerce reduced local pricing power. Yet much of the CPI relies on collection methods developed decades before online retail became commonplace. The BLS has spent years experimenting with scanner data and other new sources, but it never fully adopted those practices because of its caution and budget constraints.

During the productivity boom of the late 1990s, then-Fed Chair Alan Greenspan became deeply interested in inflation measurement because accurately pricing computers was essential for understanding productivity growth. Indeed, he testified before the Senate Finance Committee that the reported annual CPI was actually 0.5 to 1.5 percentage points higher than it should be, or, in economics-speak, "biased upward," because of unmeasured tech-related quality improvements.

Today's AI boom presents a similar challenge. Suppose a coding assistant costs the same now as it did six months ago but is twice as capable. Has inflation really remained unchanged? Or has quality improved so much that economists are understating productivity growth?

ALL THESE DEVELOPMENTS have created an opening for alternative data. The private sector controls much of the information that the modern economy

generates, including online prices, payroll records, payments and earnings-call transcripts. Companies and universities can incorporate new technologies more quickly than official statistical agencies.

For inflation, the now-inactive Billion Prices Project, founded by economists Alberto Cavallo at Harvard University and Roberto Rigobon at Massachusetts Institute of Technology, demonstrated years ago that it could replicate official inflation measures using web-scraped data. Software-maker Adobe Inc. publishes a digital price index, based on millions of online transactions. A company called Truflation has built a high-frequency inflation gauge that investors increasingly monitor alongside the official CPI. Technological advancement has made processing big data a less daunting task. Computing power has expanded dramatically with AI. Large language models can track macroeconomic signals from hundreds of earnings transcripts with an accuracy that was previously impossible.

None of these technologies eliminates the need for careful methodology, but they let economists work with data that at one time would have been too large or messy to use. At the same time, investors are constantly trying to find an edge, and timely and accurate data can make all the difference. Bloomberg Economics has

10

BLOOMBERG MARKETS

---

What Earnings Calls Say About the Economy

Growth optimism and inflation concern in S&P 500 earnings calls

Transcript-based index of growth optimism

Transcript-based index of inflation concern

Growth optimism index is a measure of the pervasiveness of industries expressing growth optimism in their earnings calls. Higher values mean more industries reporting expansion signals, such as guidance raises, record orders or margin improvements. Inflation concern index is a two-quarter moving average tracking the breadth of inflation-related discussion in S&P 500 earnings calls, covering wage pressure, pricing power, consumer demand and supply costs.

Source: Bloomberg Economics

approached this opportunity through the Bloomberg Price Project, an effort to build a consumer basket from the ground up while remaining faithful to the official BLS methodology. The project tracks roughly 140,000 products and more than 321,000 monthly price observations, about three times the number collected for the official CPI.

That granularity provides a richer picture of inflation dynamics. At any moment, hundreds of CPI components are responding to different supply-and-demand shocks. Looking beneath the published aggregates reveals developments that are still hidden inside headline inflation numbers. One interesting data point has been the sharp acceleration of prices for hard drives, SD cards and memory modules, as a result of the AI build-out.

Alternative datasets also let economists ask questions that elude official statistics, illustrating the complementary strengths of public and private data. Government agencies provide rigor, consistency and transparency. Private companies offer flexibility, detail and the freedom to explore new questions.

The BLS employs roughly 2,000 people, many of whom collect, validate and review economic data. AI agents can perform portions of the initial screening and validation traditionally carried out by field staff. AI is unlikely to

replace official statisticians, but it can augment their work, stretching scarce public resources while improving speed and consistency.

SOMETIMES THE MOST valuable information comes from what companies say before developments appear in the official data. For decades the Fed has relied on the so-called Beige Book, anecdotal reports from businesses across the country, to supplement traditional economic indicators. Former Chair Jerome Powell frequently referred to those reports during policy debates because they could provide something of an early peek at economic shifts.

Bloomberg Economics has taken a similar approach through what we call the Orange Book. It uses AI to analyze thousands of corporate earnings-call transcripts for recurring themes involving hiring, pricing, investment, consumer demand and inflation.

The value of that approach has become particularly clear during the Iran war. Historically a sharp rise in oil prices could damage US growth. Yet the Orange Book suggested the economy had strong momentum in the first quarter. Across industries, executives continued describing healthy demand, accelerating defense orders and expanding AI investment. They projected more confidence than traditional macroeconomic models

would have implied, and the data kept us from reflexively predicting a recession.

Executives also recounted how AI was changing work itself. Some companies reported producing more with fewer employees. Others emphasized these kinds of productivity gains rather than outright job reductions. Whether those changes will translate into weaker employment remains uncertain. But earnings calls provide an early window into how businesses are adapting.

The US still has the world's most sophisticated statistical system, and government data will continue to anchor economic analysis. But the craft of market economics is changing.

When I began my career more than two decades ago, economists and traders focused on the calendar of publicly available data to forecast market-moving monthly releases such as payrolls, the CPI and retail sales. Today the market devotes more resources to finding macro signals from higher-frequency and esoteric data that can separate you from the pack.

The future of market economics doesn't lie with Washington or Wall Street alone. Government agencies set the standard; the private sector expands the frontier. ●

Wong is chief US economist for Bloomberg Economics.

VOLUME 35 / ISSUE 4

11

---

The Curse of A Strong Currency

By VINÍCIUS ANDRADE and NICOLLE YAPUR

PHOTOGRAPHS BY JUAN CRISTÓBAL COBO

IN COLOMBIA, where some of the world's most coveted coffee grows on the steep slopes of the Andes, the biggest threat to the fortunes of farmers these days isn't drought, disease or low prices. It's a strong currency.

With coffee prices high, Edmy Yojana Correa, a third-generation coffee farmer, had been earning enough to buy agricultural equipment and replace her car. But this year, through mid-July, the Colombian peso has rallied 17% against the US dollar, eating away at her profits. Every buck paid by her customers abroad is now worth less at home, leaving her short of money for wages, fertilizer and other supplies.

Across almost every corner of Latin America, exporters are facing the same squeeze: Local currencies are strengthening against the greenback, the medium of exchange in which most commodities are priced and traded. From Mexico, where the peso trades around the clock, to Paraguay, where the market for the guarani is relatively illiquid, the phenomenon is rippling across ►

Santiago Silva at his coffee farm in Supatá

12

---

---

Cortes and Silva on their farm about two hours from Bogotá

industries, households and central banks.

In Colombia, it's showing up in flower crops as well as in coffee fields. Correa's earnings are down 40%, so she's considering borrowing more to stay afloat. "We started making improvements," she says, from her small farm some 6,000 feet above sea level in the mountains of Huila, a premier coffee region. "That helped us, but right now, we are starting to falter." In Costa Rica, the colón is so strong that banana farms are closing down. And for a Paraguayan

snackmaker, the state of the guarani is coming up at board meetings.

Since the start of 2025, a basket of Latin American currencies has appreciated on average 19% against the greenback. Although a weaker dollar lifted emerging markets more broadly, the gains in the region have outpaced those in other parts of the developing world. They reflect, in large part, the flip side of a dollar that softened in the early months of Donald Trump's second presidency as investors grew wary of his trade policies and sought to allocate more money to the rest of the world.

In other ways, the resurgence shows how deeply undervalued much of the region had become. For years, Wall Street's fascination with the US stock market boom and artificial intelligence left little room for Latin America. "Periods of strength have been rare" for the region, says Phoenix Kalen, global head of emerging-markets research at Societe Generale in London.

Weak currencies fueled higher prices and undermined confidence. Policymakers in Latin America's largest economies raised interest rates to some of the highest levels in the world to quell inflation. That turned them into targets for carry traders, who

borrow in places with lower rates, such as Japan or Switzerland, and invest where yields are higher. Those same rate differentials also made it more expensive and risky to bet against Latin American currencies.

Several smaller economies, meanwhile, pursued policies aimed at deepening their capital markets and attracting investment, a combination that tends to support currencies over time. Uruguay has been seeking to gradually reduce its reliance on the US dollar, while Paraguay implemented economic reforms that culminated in the coveted investment-grade rating.

Yet few policymakers anticipated the magnitude of the rallies that would follow. Earlier this year, Uruguay's central bank cut interest rates, citing "disruptive dynamics" after inflation undershot its target for several consecutive months. In Mexico, a resilient peso was part of the reason why its central bank continued to lower borrowing costs this year.

The arrival of a new Federal Reserve chair has tempered expectations for further US dollar weakness. But that doesn't mean emerging-market currencies are set for a bear market, according to Morgan Stanley strategists. In fact, analysts at Standard

Through the Roof

Spot return of selected Latin American currencies versus the US dollar, Dec. 31, 2024, through July 15, 2026

Colombian peso	36.8%
Paraguayan guarani	29.1
Brazilian real	21.6
Mexican peso	19.8
Costa Rican colón	13.3
Peruvian sol	10.5
Uruguayan peso	9.4
Chilean peso	7.5

Source: (WCRS+GO+)

14

BLOOMBERG MARKETS

---

Chartered Plc and Societe Generale expect Latin American currencies to remain relatively resilient over the medium term.

Part of the appeal is that the region has advantages compared with other emerging markets. Higher energy prices threaten many Asian economies that import fuel, but several of Latin America's largest economies stand to benefit as energy exporters. The region's relative distance from today's geopolitical flash points has also become more attractive to investors.

To be sure, some industries are benefiting. Take Brazil. Last year the real jumped more than 12% against the dollar. That means it suddenly became cheaper for a family in São Paulo to travel to the US. Trips to Orlando, long a favorite destination for Brazilian tourists seeking theme parks and shopping, rose 5.6%. The year prior, when the real was falling, travel to the vacation spot was almost unchanged, according to Visit Orlando.

Residents in neighboring Argentina are also traveling more. The peso has appreciated dramatically in real terms thanks to President Javier Milei's sweeping economic overhaul, which slashed the fiscal deficit, tightened monetary policy and dismantled some capital controls.

But businesses are generally more

concerned about the growing downside. Even cheaper foreign travel means tourists to Latin America can get discouraged. In Argentina, residents in 2025 spent more than twice as much on trips abroad as foreign visitors did in the country.

In Costa Rica, Del Monte Corp. announced in May that it was closing four banana farms, resulting in 850 workers being laid off. The company cited the sharp rise in the colón—up 13% since the start of 2025—as well as rising production costs. It said the foreign exchange rate has a direct impact on exporters.

It's a problem, too, for CMA Paraguay SA, which produces about 230 metric tons of potato chips a month that are shipped across South America. Founded in 2018, the company rarely discussed foreign exchange rates in previous years, even though exports account for about 90% of its sales. Now, Florencia Fustagno, a board member, says it's a weekly topic at the panel's gatherings. "We thought it was something temporary, but now that a year has gone by, we believe it's going to stabilize" around the current exchange rate, she says. "We'll have to adapt to this situation and find a way to adjust our cost structure."

Few places, though, are seeing more

consequences from currency appreciation than Colombia, where the peso is near its strongest level in seven years. The finance ministry took advantage of the situation to scoop up dollars and unwind a swap transaction in Swiss francs.

But a couple of hours outside Bogotá, the capital, the strong peso is interfering with the retirement plan of Santiago Silva and Sonia Cortes. The couple saved for more than two decades to buy their own farm. Five years ago, they finally found a spot about 6,500 feet above sea level, surrounded by native forest and a stunning view of the Eastern Andes.

On the humid, rain-soaked hillsides, coffee grows among orchids and exotic trees, with only a pack of rambunctious dogs disturbing the calm. Silva and Cortes planted coffee prized by global buyers and built their own drying facility. Quartz crystals in the soil filter the water that nourishes the crop. Years of preparation culminated in their first harvest last year.

Silva's revenue is down about 20% since then. He's focused on improving the quality of the arabica and robusta beans he's growing, which will fetch more on the international market and make up some of the foreign exchange differential. "The ►

Freshly harvested coffee cherries at Cortes and Silva's farm

---

Silva and Cortes' farm in Supatā

16

BLOOMBERG MARKETS

---

character of Colombian coffee comes from the people,” Silva says, explaining why he’s betting on its enduring appeal. Others, he fears, will not be able to adapt and will give up on coffee for other crops.

The couple have jobs in the city they can fall back on. Many of Colombia’s growers don’t. Small family farms—most only 1 to 3 acres, where coffee cherries are picked by hand—dominate the country’s industry. In January, President Gustavo Petro raised the nation’s minimum wage, increasing labor costs just as the peso strengthened.

The National Federation of Coffee Growers exports about one-fifth of Colombia’s premium coffee and guarantees a market for all coffee produced in the country, helping cushion swings in global bean prices and establishing benchmarks.

Exporters warned in early July that the peso’s rapid appreciation is eroding Colombia’s competitiveness and urged the government to expand hedging tools and credit lines. “Coffee growers are an export-oriented sector, and in that sense they are entirely dependent on the current exchange rate,” says Esteban Ordoñez, the federation’s chief commercial officer.

Back at Correa’s small farm in Huila, the main harvest season is still a couple of months out. The weather has been finicky, but she expects small blossoms will soon start showing up on the 7,500 or so trees that dot the rolling hillsides here.

For most of the year, she and her husband tend to the trees themselves. But, as the parents of two small children, they’ll hire a few pickers when the cherries begin to ripen. The region is known for producing high-quality arabica coffees; caturra and pink bourbon varieties are valued for their sweetness and winelike acidity.

Some of Correa’s beans will end up in Nestlé SA’s Nespresso coffee pods. She has avoided harsh chemicals, earning Rainforest Alliance certification for her sustainable practices. The certification brings an added premium, but it also comes with higher costs, including organic fertilizers. To cover those expenses, the family has already had to borrow money. “You can’t neglect the crop,” she says. “The costs are fixed. The only thing that’s variable is the profit.” ● —With Ken Parks

Andrade, in New York, and Yapur, in Bogotá, cover emerging markets.

VOLUME 35 / ISSUE 4

17

---

Markets

Asia's Small-Fry Investors Are Hitting the Big Time

By ALEX GABRIEL SIMON and WINNIE HSU

ILLUSTRATION BY CHAU LUONG

Home Cooking

Cumulative fund flows in selected Asian markets*

✓ Domestic ✓ Foreign

Funds based in nine selected Asian markets invested $635B in their respective home markets from 2018 through 2025

*Flows are from EPFR-tracked funds in China, India, Indonesia, Malaysia, the Philippines, South Korea, Taiwan, Thailand and Vietnam. Excludes 2018 domestic flows data for Indonesia, which was unavailable. Source: Bloomberg analysis of EPFR data

HOMEGROWN INVESTORS are overturning the balance of power in Asia's markets. For decades investors in New York and London, primarily big institutions, drove trading volumes in Asia. But the continent's rising middle class is shifting its saving from gold and real estate toward stocks, bonds and other financial assets, becoming the primary source of capital in Asia. The increase in disposable incomes helps too.

In some cases, governments are encouraging households to take more risk. India, Japan and South Korea are pushing to improve corporate governance and boost shareholder returns to broaden the appeal of stock investing. Families are taking the bait, funneling savings into equities through mutual and exchange-traded funds as well as individual stocks. "You have this budding equity culture," says Alexander Redman,

chief equity strategist at CLSA in Singapore.

The shift has broad implications because Asia accounts for 56% of the world's listed companies—shares valued at $43.5 trillion, including those at the heart of the artificial intelligence boom. The growing prominence of domestic investors means markets in India, Taiwan, Vietnam and elsewhere have become less sensitive to the whims of overseas traders who like to yank their money in times of crisis, such as the Iran war. "Foreign flows can be very fickle," Redman says. (Flows are purchases minus redemptions.)

Yet the change carries its own risks, raising concerns about overheating valuations and speculative trading. Domestic investors could further concentrate capital in local champions and popular sectors, potentially leaving markets more vulnerable

to abrupt reversals. "Once sentiment turns, everyone may try to exit through the same narrow door," says Charu Chanana, chief investment strategist at Saxo Markets. For example, Chinese stocks slumped in January after authorities tightened rules on margin financing. The move signaled an unease over an AI-led rally that had sent benchmarks to multiyear highs and driven turnover to record levels.

It's hard to overstate the magnitude of the shift to domestic investing. Funds based in emerging Asian economies poured $557 billion into their stock markets from 2022 to 2025, more than seven times the sum they invested during the previous four years, according to financial data provider EPFR. By comparison, the amount that foreign-based funds committed dropped 81%, to $26 billion.

18

BLOOMBERG MARKETS

---

Standing on Their Own

Foreign flows and stock performance in selected Asian markets in 2025

Market	Foreign flows	Increase in benchmark stock index
South Korea	▼ $4.5B	76%
Vietnam	▼ $4.8B	41
Taiwan	▼ $7.8B	26
Indonesia	▼ $1.1B	22
India	▼ $18.8B	11
Malaysia	▼ $5.2B	2

Source: Bloomberg

Rank-and-file investors will only grow in clout. The Organization for Economic Cooperation and Development projects 65% of the global middle class will live in Asia by 2030. The region's share of global private wealth climbed from 6% to 21% in 25 years, according to a 2025 United Overseas Bank Ltd. report. By 2029, Asia could claim a quarter of that stash, worth $99 trillion. "When I came to Asia in 1999, it was capital-constrained," says William Bratton, head of cash equity research for

the Asia-Pacific region at BNP Paribas. "Domestic capital pools were relatively small, and if you wanted to do something, there was a substantial need to access foreign capital."

Consider other striking data points: India in the past six years has overtaken the UK and France to become the world's fifth-largest stock market by the value of its shares. At 14.5 million, the number of securities accounts in Taiwan roughly equals 62% of its population. Domestic investors

commanded a 70% share of trading in Indonesia last year, while in Vietnam the proportion of foreign investors fell to about 14% as of July 23, an all-time low, HSC Strategic Market Research found.

In each of these markets, as well as some others, benchmark stock indexes rallied in 2025 even as foreign investors pulled their money. When Asian markets open, the stock purchases of local investors are "setting the tone before Europe comes online," says Rajiv Batra, head of Asia and co-head of global emerging-markets equity strategy at JPMorgan Chase & Co. in Singapore.

Overseas selling can still hit Asia markets hard. Indian and Indonesian shares struggled this year as foreign investors pulled back because of higher oil prices, weaker currencies and few domestic AI-linked stocks. Still, without local buyers, Batra says, those markets "would have gone to the cleaners." • —With Cindy Wang, Prima Wirayani and Nguyen Kieu Giang

Simon covers equities from Mumbai; Hsu, from Hong Kong.

---

Prices

Japan's Inflation Paradox

By YOSHIAKI NOHARA and SAKURA MURAKAMI

PHOTOGRAPH BY NORIKO HAYASHI

A FIVE-HOUR DRIVE north of Tokyo, Chieko Sugai lives alone in a two-story house that's seen better days. The paper sliding doors are torn, and spiderwebs lace the corners of empty rooms.

In the coastal city of Murakami, the 72-year-old widow relies on a monthly pension of 110,000 yen ($671). In April kerosene costs alone ate up more than a third of her budget, because of the Iran war. "When I don't have money, I just have to put up with it because I have no other choice," she says.

With many senior citizens facing a similar situation in her community, Murakami passed a petition in December, urging Prime Minister Sanae Takaichi to boost pensions in line with inflation to

"revitalize local towns." In 2024, Japanese age 65 to 74 and living alone found themselves digging a deep financial hole. Each month, on average, they were spending 21% more than their disposable income of 142,000 yen, according to the latest government data.

The rising cost of living is challenging Takaichi's triumphant rhetoric about Japan's economy. It's something of a paradox: Inflation has caused pain in many quarters, while also ushering in an era of renewed dynamism. Company profits are growing. Workers are getting the strongest wage gains in decades. And the stock market is surging, thanks in part to the artificial intelligence boom, with the Nikkei 225 more than

doubling since US President Donald Trump's initial tariff shock in April 2025. "Japan is back," Takaichi declared during a dinner at the White House in March.

While inflation is bedeviling countries around the world, including the US, Japan represents a special case. For three decades the primary concern was deflation. A succession of prime ministers and central bankers unleashed record stimulus after the bursting of stock market and property bubbles in the late 1980s and early '90s caused deep scars. Japan accrued public debt now worth more than twice the nation's gross domestic product, yet prices and the economy kept flatlining.

In the early 2020s the supply shocks stemming from Covid-19 lockdowns and Russia's invasion of Ukraine caused inflation to return. From mid-2021 it began accelerating to peak at 4.3% in early 2023. It was well below the post-pandemic highs of 9.1%, 10.6% and 11.1% in the US, Europe and the UK, respectively, but a shock for a society unaccustomed to rising prices.

Now inflation is back to just below the central bank's 2% target, and there are growing signs it's becoming more durable. Given its experience with falling prices, the government sees that level of inflation not as a potential threat but as a fragile flame to be nurtured.

But even at this level, rising prices are stoking divisions. Tokyo's young and educated workers are winning pay gains that exceed inflation, and its well-heeled residents are benefiting as the equity and housing markets surge. Pensioners and

Epic Journey

Japan's consumer price index (excluding fresh food), year-over-year change

Source: (JNCPIXFF GP )

20

BLOOMBERG MARKETS

---

Sugai at home in Murakami

workers on lower wages in rural areas, however, are going backward as the cost of living increases. Although support for Takaichi's premiership remains high following a resounding electoral victory that saw her win a supermajority in the lower house in February, opinion polls show prices remain the top public concern.

Takaichi has sought to cushion households from rising costs through measures including energy subsidies. In her election campaign she vowed to suspend the 8% sales tax on food for two years. The prime minister has said she'd do more if needed. But many pensioners have seen little relief, as their payments have lagged behind the cost of living for years. In 2026 they rose about 2%, below last year's 3.2% inflation rate.

Investors are watching to see whether Takaichi can execute her plans without worsening a debt burden that's already the largest in relation to GDP among major economies. In January long-term bond yields spiked on concern that her Liberal Democratic Party would need to ramp up spending to win the election, shaking global markets and reminding the leader of the constraints on her policy plans.

The government is encouraging its famously risk-averse citizens to invest more in the stock market as a way to achieve gains that outpace inflation.

In 2024, Japan expanded its decade-old Nippon Individual Savings Account program, which now permits individuals to sock away as much as 18 million yen over their lifetime and reap any profits, tax free. Investments totaled 71 trillion yen at year-end 2025, already exceeding the 56 trillion yen the government had been targeting by year-end 2027, according to Japan's Financial Services Agency. Household stock holdings stood at 398 trillion yen at the end of March, a 29% increase from a year earlier, while cash and deposits rose just 0.6%, to 1,126 trillion yen, according to the Bank of Japan.

On a recent Saturday a nascent rise in risk-taking was on full display during an investment expo at the "Tokyo Big Sight" convention center. Exhibitors touted stocks, real estate, gold, whisky and antique samurai swords. A makeshift bar sold cocktails dubbed "Inflation," "Warren Buffett" and "Margin Call." One presenter warned about inflation evaporating cash right out of people's wallets. An influencer with her own YouTube channel inspired the crowd with a

story of quitting a corporate job to achieve financial freedom through investing.

Yukiko and Hidekazu Kamino stood at the front of the line to hear her. The couple—married for about 25 years and with two kids—had traveled from Fukushima prefecture to attend the expo. Hidekazu started buying financial products a few years ago, breaking from his parents' teaching of parking money in bank accounts and life insurance policies. He bought US stocks as he saw brighter prospects for the country, a move that paid off as the yen's slide to its weakest level in four decades boosted his American shares' value.

Even though their household is feeling the pinch of higher prices, the Kaminos aren't sold on Takaichi's plan to stall the sales tax on food. "It's certainly popular, and it may be a lifesaver for pensioners," says Hidekazu, 58. "But it's bound to lead to tax hikes."

Back in Murakami, a city of 52,000 known for its sake breweries and stunning sunsets, residents say they need relief. More than 40% are 65 or older. Sugai and her husband moved there in 2019 to retire after years working for a local rice cracker company. But hardships soon followed. Her husband developed cancer and dementia and died in January 2025. Paralysis in her left leg from a stroke forced her to give up driving. "My pensions won't rise, and I worry they might go down," she says.

Unable to drive and reluctant to spend 10,000 yen on round-trip taxi fare to the nearest supermarket, Sugai relies on weekly visits from a food truck for her groceries. Its operator, Ayumi Kato, makes similar stops in the isolated communities scattered between rice paddies and empty houses across the region, driving past leftover campaign posters featuring Takaichi's promises to make all of Japan "Strong and Prosperous." Kato says she runs the business to make a living but also wants to help the vulnerable, adding only a small margin of about 20 yen over supermarket prices to cover her fuel costs.

"People are really being selective now. Instead of buying three snacks, they'll just get one," Kato says. "I really have to think it through when setting prices. Profit is important, but everyone here is elderly, and some of them are really struggling." ●

Nohara and Murakami report on economics and politics from Tokyo.

VOLUME 35 / ISSUE 4

21

---

Education

Colleges Bank On an AI Training Boom

By LIAM KNOX

ILLUSTRATION BY VINCENT KILBRIDE

FOUR YEARS AGO, Frederick Duff felt his career was stagnating. A mortgage broker in southern Ohio, he'd grown restless and wanted a change. "I was in the mortgage industry forever," he says. "I could do it drunk, upside down in a swimming pool."

Duff, 57, thought about going back to school. But he already had his MBA, and he didn't want to spend the time or money on law school. Instead, he enrolled in a non-degree online program to learn about a quickly emerging technology that he believed would be more relevant to the workforce: artificial intelligence.

It was a good hunch. The number of employers who expect entry-level workers to be AI-competent has nearly tripled since last fall, according to the National Association of Colleges and Employers. Workers' anxiety about AI has surged, along with demand for quick training programs. In 2022, before the launch of ChatGPT, just 2% of the professional certificate market was focused on AI. By 2026, short-term AI training programs—often called boot camps—had exploded to nearly a third of the credential market, according to workforce database Revelio Labs.

Policymakers in Washington are encouraging schools and companies to launch more programs as part of a bid to boost workforce training. Some employers are even willing to foot the bill to prepare their workers for what they see as the fast-approaching future. And colleges are scrambling to meet the moment, betting on short-term credentials—and AI in particular—to shore up flagging revenue.

The surge recalls the early 2010s coding boot camp boom. With traditional universities moving at a glacial pace to set up computer science departments, credential programs stepped in. "Learn to code" became the panacea for widespread unemployment following the 2007-09 recession, and students rushed to fill seats. Many enrolled in for-profit programs that sprang up, though the quality varied widely and regulators accused some schools of misleading students about how much they could make after earning their credentials.

Today many traditional schools are offering training in AI tools for business strategy, deep dives into machine learning and crash courses in large language models. And they're marketing to everyone from C-suite executives and software engineers to the AI layman looking for a leg up in the job market. For these schools, the pressure is on, not just to prepare students for employment, but also to find fresh revenue streams.

Universities face declining enrollments and threats to their business model; international students are turning elsewhere; high-cost graduate programs are at risk; and billions of dollars in federal research funding are suddenly up in the air. Certificate programs can help fill the gap. There are nearly twice as many of them as there are degree programs in the US, and Americans pay about $2.3 trillion a year for them, according to 2025 data from the Credential Engine. AI made up around 16% of all certificates awarded last year, up from just over 1% in 2022, according to data from the Burning Glass Institute, a think tank focused on the future of work.

Georgia Institute of Technology has rolled out several microcredentials—short, relatively inexpensive, online certificate programs—focused on AI competency in engineering, education, business management and more. Ángel Cabrera, the university's president, says its next big venture is an AI master's degree delivered entirely online. "This is definitely an area of growth," Cabrera says. "Our programs are tapping markets that were not being served."

Federal policy changes under President Donald Trump are likely to encourage more growth. A new Workforce Pell Grant opened financial aid to students enrolled in programs as short as eight weeks for the first time this year. The Department of Education earmarked $50 million each for two new grant programs this year in AI training and short-term credentials. And Under Secretary of Education Nicholas Kent says he hopes new caps on federal lending for graduate students will encourage colleges to pivot from expensive degrees to flexible credentials with more direct employment outcomes. "We've made a tremendous amount of investments as an administration in AI education for the workforce," Kent says.

HOPES OF AN EMPLOYMENT boon may be premature. Recent data shows a slowdown in hiring has driven rising unemployment among 18- to 24-year-olds more than a shortage of AI training. Erik Leiden, managing director of workforce strategy at Burning Glass, says many of the AI boot camps cropping up don't necessarily match jobs that exist yet. Some programs are built around

22

BLOOMBERG MARKETS

---

the products of major AI labs, primarily at Anthropic, Google, Microsoft and OpenAI, and may prove less relevant to jobs that use different platforms. "Everyone is telling schools, you need to get in on AI, it's urgent, it's the future, it's here," he says. "But it's tough for colleges to know what students are going to need in the workforce, because it's such an open question right now."

That's not accounting for the flood of for-profit companies advertising their own AI boot camps, including some that were players during the coding boot camp boom. Gauntlet AI, for instance, is run by Austen Allred, whose company BloomTech agreed to pay a fine to the Consumer Financial Protection Bureau, without admitting wrongdoing, to settle allegations it made

replace tuition revenue from full-time students; colleges can't charge nearly as much for a few weeks of online classes as they can for a year of on-campus education. Robert Towner, chief business officer for Stevens' College of Professional Education, says companies like Hartford send hundreds of students their way, but can end up paying as little as $2,500 per head—less than a tenth of the typical yearly cost of a master's degree. Online credentials are almost always cheaper to administer, however, and the overhead doesn't rise substantially with each new student.

Because of that, schools can compete on price as programs proliferate. Temple University offers a one-semester AI credential for $495. The self-paced program at the

7,500 students have enrolled in the programs; for context, Purdue's entire freshman class last fall was about 9,000. Peroulis, who's also an electrical and computer engineering professor at the school in West Lafayette, Indiana, says it has dozens of new AI certificates in the works.

He says the programs more than pay for themselves, but they're primarily driven by workforce needs. "A lot of employers have told us that students need to be upskilled not in a matter of a year or two, but in a matter of a week or two," Peroulis says.

Duff says his experiences with AI courses have been mostly positive. He's taken five in as many years, ranging from a nine-month intensive in machine learning at the University of Texas at Austin to a

misleading claims about job placements and program costs. "Gauntlet AI operates under a model where students never pay us anything," says spokesman Josh Martin. Leiden says the nascent AI boot camp market is a "Wild West," loosely regulated with widely varying employment outcomes. "For students, that's really scary."

AI certificates are sometimes offered in partnership with employers who sponsor the programs or cover tuition. That's the model being tested at Stevens Institute of Technology, where leaders hope demand for AI training will help them weather financial headwinds. This fall, Stevens is piloting a 12-to-16-week AI program for employees of Hartford Insurance Group Inc.

The certificate programs cannot

University of Texas at San Antonio costs only $49. The University of Maryland is handing out certificates for free.

Purdue University offers programs as short as a few days and as cheap as $99, as well as monthlong courses for as much as a few thousand dollars. Last year, Purdue's AI microcredentials became the first of their kind to be accredited by ABET, an accrediting organization focused on engineering and technology programs. Most colleges haven't reached that threshold.

Dimitrios Peroulis, Purdue's senior vice president for partnerships and online education, says the school is trying to be flexible as it expands its footprint in a fluid market. The strategy seems to be paying off. In less than two years, more than

microcredential in generative AI at Purdue. But he says the programs changed as they became more common. Some of them are little more than prompt-writing tutors. "It depends on what you want to get out of them," he says. "You have to go in with your eyes open."

Duff now teaches in the MBA program at Bellarmine University in Louisville, Kentucky. He also started his own AI consulting firm, and he's teaching an AI boot camp run by a for-profit educational technology platform. And he still keeps a hand in the mortgage business—with some help from AI. ●

Knox covers education policy for Bloomberg News in Washington.

VOLUME 35 / ISSUE 4

23

---

The Terminal

Functions for the Market

ECAN

p30

GC

p32

WSL GEOCREDIT

p38

NEWS ON ORANGE BOOK

p40

24

---

THE US CONSUMER powers the country's economy. To dig into whether spending patterns are changing, go to {ECAN US RETAIL SALES }. What might hold back US consumers from loading their online shopping carts? Paying more at the pump, for one thing, or other higher costs related to geopolitical shocks such as the Iran war. For a sample worksheet that lets you track geopolitical risk metrics from risk-intelligence provider Seerist Inc., run {WSL GEOCREDIT }. For more FFM stories, go to {FFM }.

GCDS

p35

BQNT

The Credit Spread Heat Map example project in BQuant enables you to spot areas of opportunity in a bond index.

ASKB

p40

IN AUNZ

p42

25

---

Asset Management

J. Rothschild's CEO on Its Fund's Discount and Owning SpaceX and Anthropic

By CLAIRE OBUSAN

PHOTOGRAPH BY KATE PETERS

MAGGIE FANARI TOOK the helm two years ago as chief executive officer of J. Rothschild Capital Management, which manages RIT Capital Partners Plc, one of the UK's largest investment trusts. The firm carries the name of Jacob Rothschild (1936-2024), who founded Rothschild Investment Trust in 1971 as a vehicle for the famed European banking family's English branch. After leaving N.M. Rothschild & Sons in a dispute over the direction of the investment bank in the 1980s, he listed the investment trust, renamed RIT, on the London Stock Exchange in 1988. The Rothschild family remains the largest holder of the investment trust, which has £4.6 billion ($6.1 billion) in assets.

Fanari, a 47-year-old British-Canadian, first crossed paths with the Rothschild team about 15 years ago when she headed high-conviction equities at Ontario Teachers' Pension Plan. She joined RIT's board in 2019 and was named CEO of JRCM in January 2024. Fanari also chairs the firm's investment committee. Her conversation with Bloomberg News Senior Editor Claire Obusan has been edited for clarity and length.

CLAIRE OBUSAN: What's the clearest way to explain RIT to someone who's never encountered it?

MAGGIE FANARI: Think of it as a family office that is really looking to maximize as much growth as possible while minimizing or mitigating downside risk. So equity-like returns with lower risk. How do I translate that into numbers? Since inception [in 1988], we've generated a return of 10.6% a year on average. We've captured 71% of the monthly market rises and only 40% of the market declines. You get a very nice, strong compounding portfolio.

CO: One of the hallmarks of RIT as an investment trust, with a closed-end fund structure, is that you have permanent capital. What advantages does that give you?

MF: It gives us a couple. What it allows us to do is have a multi-asset-class portfolio. We invest in public markets but also in private markets. You really need long-term permanent capital to be able to invest in private markets—be that private funds or directly in some great companies that we own.

It gives us a differentiated appeal to our partners, because

they know we never have to go to them and say we need to sell because we have fund redemptions. The other component is, when you see volatility in the markets like we have over the last couple of years, it means we don't have to sell at inopportune times. In fact, it means we can be quite tactical and buy during market downturns.

CO: How much of the challenge, when you stepped into your role, was portfolio construction, and how much was organizational?

MF: The culture was great. A lot of it was just a little bit more organizational. Also, too, our firm is publicly listed. Anyone can invest alongside us in our deal flow. But what we really needed to do was enhance our transparency with shareholders and also build out certain departments and improve our investor relations function, our communications function, and really be much more front-footed than we had in the past. We've received a lot of positive feedback from being able to do that over the last two years. I do fundamentally believe that being transparent and giving data to your shareholders is particularly important, because then they can track how the company is doing. Then the returns and how the portfolio is operating just kind of lets them say, "Yeah, I expected that. I expected them to be down that month," or, "I expect them to be up that month."

CO: RIT has traded at a meaningful discount to its net asset value for years. With the changes you've made, what do you think the discount is a function of? Is it a communications problem or something else?

MF: One thing is sector. Normally what you see is a correlation between discounts and when interest rates are high. When interest rates come in, the discounts come down. Then the other one was, a few years ago, there was some concern around our private portfolio. There's a level of transparency you can provide, but it's not the same as a public stock that gets a mark to market every day.

What we've been able to demonstrate over the last two years is we've been able to realize a third of our portfolio. And that's all been above our NAV price. Last year our direct portfolio was up 50%, our private portfolio was up 18%. What we're known for is we look to invest alongside the best private equity and venture capital general partners in the world. Thrive, Greenoaks, Iconiq, Ribbit—these are some of our core partners. Because they're early on key themes, ►

26

INSIDE THE TERMINAL

---

Fanari

INSIDE THE TERMINAL

27

---

including artificial intelligence and fintech infrastructure, we're able to realize those returns, while other people say, "The privates market is really muted. There are no realizations."

Last year we had three exits. They were all investments made in 2021. The average internal rate of return was 22%. It was close to 2.2 times return of capital and about 100% over the last, December 2024, mark. Again, we just need to continue to get our track record out and people to understand that we are different. Our brand gives us the ability to work with, in our view, some of the best partners in the world.

CO: What were the three exits?

MF: Three companies: One was called Webull [an online brokerage that went public after merging with a special purpose acquisition company], another was Xapo Bank [a Gibraltar-based private bank for Bitcoin holders that did a management buyout]. And then the other one is [data-labeling company] Scale AI, which was acquired by Meta. So you had two M&A takeouts, and then you also had one company go public.

CO: For the investors who focus on the discount, what are they missing?

MF: We do trade at a discount, as do many of our peers in the market. For new investors coming in, it's very much an opportunity. It's very much downside-protected through the discount. Last year our share price total return was up 17%, and our portfolio was up 13.5%. We will look to increasingly narrow that discount for investors, and we will continue to demonstrate strong realizations. We'll increase our transparency, and we'll continue to enhance our shareholder engagement.

CO: I have to ask about SpaceX. You caught on very early, back at Ontario Teachers' Pension Plan. What made you decide this was something worth doing?

MF: I got to know the SpaceX team during my time at Ontario Teachers'. What was really interesting to me around SpaceX was obviously the founder [Elon Musk], an incredible founder. Really a business that was very mission-driven and made the impossible possible—NASA said you can't have a reusable rocket. I thought that was amazing.

Also just the cost advantage. What sparked my interest at the time was their ability to produce satellites for a couple hundred thousand dollars. Having been a generalist investor over my career, I'd also looked at other satellite investment opportunities: The average cost was $10 million to get one satellite into orbit. When I saw the unit economics on this, combined with one of the most consequential founders of our time with a company that had reusable rockets, it was pretty clear to me from a moat perspective, their ability for growth and what they're doing would be entirely transformational.

Then in 2024 it was the first investment I made for RIT. I came into the business and realized we didn't have any SpaceX exposure, and looking at the business and the valuation and the growth over those seven years, and having spent a lot of time with the management team, it was very apparent to me that there was more growth to come. I wanted to get as much SpaceX as I could reasonably own in the portfolio from a sizing perspective, but I thought it was definitely a company that we had to own.

CO: At one point it was your eighth-largest holding, right?

MF: It would've been in our top 10 holdings across our entire portfolio, and it was our largest direct investment after they completed the $800 billion-valuation funding round. We were up multiples of money already at that time. More so with the IPO. I think SpaceX continues to be a fantastic company with a number of growth prospects. It's becoming vertically integrated: You've got the launch business, you have Starlink. Now you've got compute, and you're combining that with xAI. I continue to see a lot of long-term runway for continued compounding and growth.

CO: Another investment you guys have recently made was Anthropic. What was it about Anthropic?

MF: We've been very selective. Maybe if I could just talk a little bit about our AI themes, going back to diversification and how we think about how much exposure we'd like to have to the AI theme. We largely expressed that through our privates portfolio, because we think most of the innovative companies are happening in privates. Last year, when we were doing diligence work on both Anthropic and Databricks [a startup provider of AI and data

"I wanted to get as much SpaceX as I could reasonably own in the portfolio from a sizing perspective"

28

INSIDE THE TERMINAL

---

Fig. 1 To chart the discount at which RIT Capital Partners has traded, run (RCP LN Equity NAV ).

applications], we could see the enormous growth that was coming through those businesses.

We actually looked to sell a lot of our existing public market software exposure. We sold companies like Microsoft, GoDaddy and Salesforce. Then we said, "How do we want to invest along the AI value chain? We're going to look at the frontier models." Looking at Anthropic and the growth rates, and just generally the fact that you have a founder who is very focused on how he's developing his models, it became very clear that Dario Amodei is an "N-of-1" [one of a kind] founder. That's part of the investment thesis really, finding these N-of-1 founders and a company where we've never seen such growth rates. This company started in 2022 with $10 million of revenue. Today it's assumed to be $45 billion and just continues to grow. That to us said, in frontier models, if we're going to invest directly, we're going to do that in Anthropic.

Then we thought about where we wanted to be in the AI infrastructure base, and we were very fortunate in our ability to invest in Databricks, which continues to grow. Then we thought about how are we invested in terms of AI applications, and we've got some great names both directly and indirectly in our portfolio, and companies like Stripe and Ramp who are really benefiting from that.

CO: In your letter in the last annual report, you wrote about AI, and the other theme you talked about was macro and geopolitical uncertainty. How are you thinking about the world right now? Where do you see opportunity?

MF: We see two structural themes shaping the way we invest. The first is AI and technology, and we have a very large allocation to the US. Our portfolio overall is 50% weighted to the US and 50% rest of world. A year ago we would've been more than 60% weighted

to the US. But what we did on the public side, going back to how do we want to be diversified, is we really believe in the theme of a more multipolar world.

We're already seeing that, where countries understand that they need to invest in their own sovereignty. What does that really mean? It means countries recognize they need to focus on reindustrializing their own economies—a bit of deglobalization that's starting to come through. We've seen countries like Germany say, "OK, we're going to run a fiscal deficit." We recognize that we need to focus on our own energy security. Also, as we've seen things like the Russia-Ukraine war—countries also recognize that they need to focus on their own defense. They recognize that they need to figure out how to build their own AI infrastructure.

What we were starting to see is the potential for the start of a commodity supercycle. As a result, we started to invest in emerging markets as well.

It's exciting for active managers and the idea of diversification. It's very different than a world we saw a few years ago, where it was largely only seven stocks driving market returns. Just given how heavily weighted cap-market indexes are today toward technology, it's not necessarily clear that investors have as much diversification as they think they do.

As we look out into the world, what key risks are we concerned about? I would say the key one that we're keeping an eye on is inflation and where rates are headed. That very much has an impact on everything that we've been discussing. We'll just have to see what that looks like, but inflation is a key risk. ●

Obusan is a senior editor at Bloomberg News in New York.

INSIDE THE TERMINAL

29

---

Economics

Is US Consumer Behavior Changing? Why It Matters to the Fed

By MICHAEL BRIODY

THE US ECONOMY ENTERED 2026 with expectations that the Federal Reserve would begin cutting interest rates. Inflation appeared to be moderating, oil prices were relatively stable, and the labor market remained resilient. A series of geopolitical events, however, including war in the Middle East, quickly changed that outlook. Rising energy prices reignited inflation concerns, and markets rapidly priced out many of the expected rate cuts for this year.

One crucial question now: Have these events changed the way Americans are spending?

The data suggests the answer is not yet. Consumer spending accounts for almost 70% of US gross domestic product, making it the engine of the American economy. Every retail sales report is essentially a national report card on the health of the consumer. It can answer a key question: Are Americans still buying what they want, or have they shifted to buying only what they need?

That distinction matters because consumer behavior ultimately determines the direction of economic growth, corporate earnings, inflation and Federal Reserve policy. The Fed isn't simply forecasting inflation—it's forecasting people. Policymakers are constantly asking themselves: Will consumers continue spending? Will businesses continue hiring? Will inflation continue to moderate?

The answers to those questions determine whether interest rates remain restrictive or begin moving lower. To use the World Macroeconomic Analyzer to track retail sales, go to ECAN US RETAIL SALES (FIG. 1). Data in recent months indicates that nonstore retailers—a category that includes businesses in the electronic shopping and mail-order houses classification, such as Amazon.com Inc.—continue to post solid growth. With an 18% weight in total retail sales, nonstore retailers in May contributed almost 0.3 percentage point to the overall 0.9% month-over-month growth of the broader category. Consumers may have become more selective in what they buy, but they've hardly stopped spending.

Market expectations, though, tell a slightly different story. Consider the dot plot, projections by members of the Federal Open Market Committee about where they expect the midpoint of the federal funds target range to be in the future. Run DOTS . To add a line showing the path of rates implied by trading in fed funds futures, click the Add to Chart button and tick the box for Fed Funds Futures – Latest Value (FIG. 2). As of early July the divergence between the Fed's projections and the market-implied path was striking. To be fair, new Fed Chairman Kevin Warsh said at the June 17 FOMC

press conference that he hadn't submitted a dot at the meeting and that the Fed's presentation of forward guidance would be reviewed this year. Yet, meanwhile, the Fed's median dot plot continues to signal gradually lower policy rates over the coming years, while fed funds futures imply a higher path for interest rates.

Why the divergence? Markets appear less convinced that inflation will return to target as quickly as the Fed expects. Sticky prices, resilient wage growth, fiscal spending, tariffs and geopolitical risks have led investors to believe the Fed may not be able to deliver all the rate cuts currently projected by policymakers. In other words, markets are pricing a "higher-for-longer" rate environment.

ANOTHER USEFUL MEASURE of market sentiment is the Cboe Volatility Index, or VIX. Often referred to as Wall Street's "fear gauge," the VIX measures how much volatility investors expect over the next 30 days. Higher readings indicate greater uncertainty and a stronger demand for downside protection.

If consumers suddenly stop buying, investors will begin to worry about slowing economic growth, weaker corporate earnings and, ultimately, a cooling economy. Those concerns typically show up in a rising VIX. To chart the VIX, run VIX Index GP (FIG. 3). Elevated market anxiety is typically signaled by the index rising to the 20-to-25 range, but as of early July it was lower than that, at 16, a level associated with normal markets. While uncertainty has increased, investors haven't yet reached the level of concern normally associated with a meaningful deterioration in economic conditions.

The most important story for the US economy isn't oil prices, geopolitical conflict or even the next Federal Reserve meeting. It's the American consumer. As long as consumers continue spending, the economy has a solid foundation. But if spending begins to slow, that foundation could weaken quickly. The Fed, financial markets and investors are all watching the same data for one reason: Consumer behavior often changes before the broader economy does.

In many ways each purchase, whether it's a new appliance, a vacation or simply clicking "Add to Cart," is more than just a transaction. It's a real-time vote of confidence in the US economy. And for the Fed, those millions of daily decisions may prove to be the most important economic indicator of all. ●

Briody is team leader for the foreign exchange, eco and crypto market specialists at Bloomberg in New York.

30

INSIDE THE TERMINAL

---

Fig. 1 To dig into retail sales data, go to {ECAN US Retail Sales }.

Online sales have held up in recent months, contributing almost 0.3 percentage point to month-over-month growth in May.

Fig. 2 To see the Fed's most recent dot plot, go to {DOTS }.

Fig. 3 To chart the VIX, run {VIX Index GP }.

INSIDE THE TERMINAL

31

---

Rates

How to Gauge Stress in Treasury Auctions as US Debt Piles Up

By LIZ GOLDENBERG

Fig. 1 Go to {ECAN US GOVDEBT } to track outstanding US government securities.

AS THE US'S NATIONAL DEBT burden grows, do Treasuries remain an attractive investment?

Total US debt rose to more than $39 trillion this year. As of late June, that figure included $31.7 trillion of debt held by the public—that is, by entities that aren't the government itself: individuals, corporations, foreign countries and so forth. The remaining $7.7 trillion was intragovernmental debt, such as bonds held by the Social Security trust fund.

Market watchers often focus on debt held by the public, because it represents the government's actual borrowings in capital markets. (Data on debt held by the public is reported daily by the Treasury Department; run {DEBTPUBL Index GP MAX } for a chart.) The Congressional Budget Office projects this component of the total debt will reach 101% of US gross domestic product this year and rise to 120% in the next 10 years, well above the record of 106% posted in 1946, in the wake of World War II.

Bloomberg has added multiple data series and tools to help market participants monitor the supply of—and demand for—US sovereign debt.

FIRST, TO CHART outstanding debt by marketable and nonmarketable securities—the latter category includes Government Account Series instruments such as those purchased by the Social Security trust fund, as well as savings bonds that are held by individuals and don't trade on secondary markets—use the World Macroeconomic Analyzer. Type "world macroeconomic analyzer" in the command line of a Bloomberg screen and click on the ECAN match in autocomplete. Next, click into the field in the upper left of the screen, type "debt" and click on the US Government Debt Outstanding match. The shortcut is {ECAN US GOVDEBT }. To show the outstanding amounts of bills, notes, bonds and Treasury Inflation-Protected Securities (TIPS), click on the white focus icon to the left of Marketable securities in the Data Table in the lower part of the screen (FIG. 1).

The International Monetary Fund tracks the ratio of total government debt (including intragovernmental obligations) to GDP for countries around the world. Use the World Countries Debt Monitor to compare total debt levels. Type "debt monitor" and click on the WCDM match. The shortcut is {WCDM }. Use the GDP drop-down to choose Debt as % GDP if it isn't already ▶

32

INSIDE THE TERMINAL

---

Fig. 2 To compare debt levels in countries around the world, go to {WCDM .

Fig. 3 To graph the Treasury Actives Curve, go to {GC I2S }.

Fig. 4 For a sample worksheet tracking Treasury auctions, go to {WSL USTAUCT }.

INSIDE THE TERMINAL

33

---

Fig. 5 To view the T-Bills Curve, go to {GC MM48 .

selected (FIG. 2). Click on the column heading to sort the list. The US's total-debt-to-GDP ratio of 120% puts the country behind only Japan, Singapore, Greece and Italy among developed nations.

Click on the figure for the US to open a chart. In December 2020, during the Covid-19 crisis, the US's total debt ratio reached a record 132%.

COUPLE THE RISING LEVELS of debt with speculation that more Treasuries will have to be sold to fund tax cuts and restock military supplies depleted by actions in Iran, and investors may be ready to punish the US government. There are multiple measures of demand for debt, ranging from the bid-to-cover auction metrics, which indicate how many times oversubscribed a specific auction was, to the shape of the yield curve.

Treasury yields rose as the Federal Reserve began hiking rates in 2022. In the past two years, the yield curve steepened following a period of inversion, when yields had been higher for shorter maturities than for some points farther out on the curve. Use the Graph Curves function to compare the curve on different dates. For the Treasury Actives Curve, run {GC I25 } (FIG. 3). You can choose specific or relative dates for your comparison.

Meanwhile, demand for new Treasuries has been concentrated in the short end—among Treasury bills and short-coupon maturities—as investors express concerns about the amount of debt and the path for economic growth in the US.

To track data on the current week's Treasury auctions, use the UST Auction Monitor sample worksheet. Type "sample worksheet" and click on the WSL match. In the amber field, type "Treasury auction" and press . Click on the worksheet titled UST Auction Monitor. The shortcut is {WSL USTAUCT }. Organized by date, the worksheet shows auction results for

recently concluded sales and the coming schedule. It identifies the instrument being sold—bills, notes, TIPS and so forth. Use the scroll bar to move to the right. Results of auctions are shown to the right of the first white dividers (FIG. 4). The section displays the high yield, the level of bid at which the sale cut off, as well as measures of demand such as the bid-to-cover ratio.

EACH WEEK THERE'S a Treasury bill sale in addition to other auctions. How much the government raises is a data point that investors track, as it affects both total supply and the supply of specific issues. Use {ALLX USTN } to see data on the Treasury's coming maturities and paydowns, as well as how much cash will be raised each week.

To see auction demand history by type of buyer, run {ECAN US TREASURY ALLOTMENT }. The chart shows how the investment funds have become increasingly important for funding the US government.

If investors are unsure of the Federal Reserve's interest-rate path or the outlook for economic growth, they're likely to put cash into the money markets—the short end of the curve—and avoid the longer-term commitment of notes and bonds.

Amid an influx of cash, money markets have been volatile. To view and compare the T-bills curve, run {GC MM48 } (FIG. 5). At the end of the second quarter, the curve was upward-sloping—indicating an investor preference for very short-term maturities—in contrast with the comparatively flat curve at the end of the first quarter. As US economic data was surprisingly stronger amid faster inflation, traders started to pull back on expectations the Fed would ease in 2026, which contributed to the movement of the curve. ●

Goldenberg is a market specialist for money markets, securities finance and fixed-income liquidity at Bloomberg in New York.

34

INSIDE THE TERMINAL

---

Fixed Income

Analyzing the Iran War Shock to the Region's Bond Market

By EBRU BOYSAN and BASSEL MAHMOUD

Fig. 1 To track trends in regional G3 debt issuance, go to {LEAG @MENAG3BONDS }.

THE RESILIENCE OF MIDDLE EAST debt markets to the conflict in Iran may provide opportunities for investors.

Regional sales of bonds denominated in US dollars, euros or yen rose 22% this year, led by Gulf Cooperation Council sovereigns and banks. Sovereign credit-default swap costs have largely fallen to prewar levels, as oil prices stay above fiscal breakevens for producers. Yet GCC bank yield premiums are still wide on concern over economic disruption. What's more, the escalation of the conflict in July has Gulf nations seeking more funding for infrastructure to bypass choke points in the region.

Use Bloomberg's LEAG, GCDS, BI, FIW, BQNT and ASKB tools for analysis.

FIRST, TO KEEP UP with developments in Iran, go to {NI IRAN } and click on AI Summary. To monitor Middle East and North African debt issuance, type "league tables" in the command line of a Bloomberg Terminal screen and click on the LEAG match in auto-complete. Click All Official Bond & Note Tables and scroll to select MENA G3 Currency Bonds & Sukuk in the EMEA section. The shortcut is {LEAG @MENAG3BONDS }. Set Role to Issuer.

MENA bond & sukuk G3 currency issuance rose 22% this year as of late July, to $114 billion. Click the chart symbol next to total. Issuance this year got off to a strong start, slowed in March and revived in April. Click the radio button to select Historical Issuance (FIG. 1). Last year's $171 billion record is within reach.

"Just to speak of resilience, we've had issuance across every country grow," Basel Al-Waqayan, a Bloomberg Intelligence credit analyst, said in a July 7 seminar. Bonds in Saudi Arabia, the United Arab Emirates and Qatar are cheap relative to the broader emerging-markets aggregate, while Kuwait and Bahrain have even more potential depending on how the conflict resolves, he said.

To check spreads on credit-default swaps, contracts that let buyers or sellers protect against or speculate on the default of a debt issuer, type "global CDS chart" in the command line and select GCDS. The shortcut is {GCDS }. Click the pencils to set the filters to All Sources, Middle East/Africa, All Ratings and Government. Hit Close. Tick Abu Dhabi, Dubai, Kuwait, Oman, Qatar, Saudi Arabia and UAE (FIG. 2). (Note: Iraq and Bahrain are both more than 200 basis points.)

The CDS fell from their highs in March, with Saudi Arabia and Oman lower than a year ago. Bond risk for Dubai is still higher, ►

INSIDE THE TERMINAL

35

---

Fig. 2 To check CDS spreads on MENA issuers, go to {GCDS }.

Fig. 3 To chart breakeven prices for regional oil producers, go to the Bloomberg Intelligence Integrated Oils dashboard at {BI INTOG }.

Fig. 4 To find GCC bonds that offer attractive yield premiums, go to {FIW @GCCBANK }.

36

INSIDE THE TERMINAL

---

Fig. 5 For the Credit Spread Heatmap example project, run {BQNT HELP } for the BQuant help page and click on Examples.

suggesting more concern over real estate and tourism than energy sectors.

Go to World Bond Markets by running {WB }, click EMEA, select YTD and rank by Range. Saudi Arabia's 10-year yield stood near 5.43% as of July 24, versus this year's low of 4.75%, while Qatar yielded 4.96%, versus a low of 4.14%. Qatar's liquefied natural gas export recovery is lagging that of crude.

In the meantime, oil prices may hold above fiscal breakeven levels, the prices at which government budgets balance. Type "BI integrated oil" in the command line and select BI INTOG for the Bloomberg Intelligence Integrated Oils dashboard. The shortcut is {BI INTOG }. Click Macro under Data Library. Click the MENA Breakeven Prices tab, then the Breakeven subtab. Click the chart symbol next to Bahrain. Click Select Additional Data. Tick Arab neighbors. Set Axis to Single (FIG. 3).

Fiscal breakeven prices for Saudi Arabia, Kuwait and Iraq as of July 24 were below the Brent price of $96. The breakeven prices were much lower for Qatar, UAE and Oman. Only Bahrain looks troubled by this metric. BI analyst Salih Yilmaz writes that as the war drags on, there is a "frozen conflict" scenario of $90-$110 oil.

TO FIND GCC BONDS with the fattest yield premiums, type "fixed income worksheet" in the command line and select FIW. Type "GCC credit" in the ticker box and select the I23117US Index match for the Bloomberg GCC Credit + HY Index: USD. Set Group By to Sector. Click Show Facets and Reset Facets. Click the gray Chart tab. Set the Y Axis amber box to G-Spread and X Axis to Bid Workout. Click Curves and Groups. In the Added Curves amber box, type "a+" and select USD US Financials A+, A, A- BVAL Yield Curve. Click the icon for that curve and All Debt. Right-click on the

chart to show points. Repeat to select Collapse all groups. Right-click on the relatively high Financials cross to expand that group. The shortcut is {FIW @GCCBANK } (FIG. 4). High-quality GCC banks still have elevated credit spreads compared with regional dollar and single-A US financial issuers.

Use BQuant Desktop, Bloomberg's sandboxed Python environment, for a credit spread heat map. Type "BQNT" in the command line and select the BQNT HELP - BQUANT: Guides and Documentation match. Click on Examples. On the left, tick Fixed Income for Asset Class. Click Credit Spread Heatmap and Add to My Projects. Run {BQNT } and double-click this notebook in your library. Click Launch Project. Click Run at top left and Run All Cells. Scroll to the bottom and set Index to EMUSTRUU Index. The average emerging-markets index spread in the 0-3 maturity bin was 53 in the AA rating segment.

Set the Index field to I23117US Index and hit the green Go button (FIG. 5). The premium for 0-3 AA GCC bonds is 85, roughly 30 points higher than for the emerging-markets index. The A segment is similar.

Finally, use ASKB by Bloomberg AI for a BI research summary. Run {ASKB }. In the question box, paste this prompt: "Summarize in 500 words what Bloomberg Intelligence wrote in June-July about GCC bank credit fundamentals. Does the backdrop suggest buying opportunities?" BI flags some names to watch as rate increases may improve net interest margins for regional banks and construction demand adjusts rather than stalls. It notes perpetual bonds offer room for spread compression if macro fundamentals hold.

Boysan is a credit market specialist, and Mahmoud is a senior technical account manager, both at Bloomberg in Dubai.

INSIDE THE TERMINAL

37

---

Fixed Income

Incorporate Geopolitical Risk in Your Analysis

By LIZ GOLDENBERG and LERA SHUMAYLOVA

Fig. 1 For a sample worksheet that incorporates Seerist data for credit analysis, go to {WSL GEOCREDIT .

GEOPOLITICAL RISK SEEMS to be everywhere these days. Whether it's political instability, cybercrime threats, government integrity or regulatory uncertainty, worries are on the rise. Such risks, however, can be difficult to quantify, let alone incorporate into your analysis of bonds or stocks.

To help with this challenge, Bloomberg has collaborated with Seerist Inc., a risk-intelligence provider in Reston, Virginia, to offer geopolitical risk scores and ratings for the 245 countries and territories that Bloomberg mapped to 7 million companies. The scores combine real-time global monitoring, artificial-intelligence-powered risk detection, verified events and human analysis to provide comprehensive threat and risk intelligence.

Real-time geopolitical risk data enables investors to make smarter decisions when comparing sovereign risk. In addition, it can help you identify when a structural baseline has broken or when a long-term relationship has eroded. The ratings are grouped into five major categories: political, operational, security, cyber and maritime. This breakdown allows investors and analysts to distinguish between permanent structural changes and transitory or event-driven disruptions.

For a sample worksheet that incorporates Seerist data for

credit analysis, type "worksheet sample library" in the command line of a Bloomberg Terminal screen and click on the WSL match in autocomplete. Type "geopolitical" in the amber field and press . Then click on the worksheet titled "Geopolitical Credit Risk Monitor." The shortcut is {WSL GEOCREDIT } (FIG. 1).

The sheet displays market-implied credit metrics and Seerist metrics such as Pulse score, which aims to take a geopolitical temperature at the country level, based on long-term trends associated with security, political and economic risks and short-term fluctuations derived from news. Use the scroll bar at the bottom of the worksheet to explore the categories. The sample sheet comes loaded with sovereign bonds from issuers across regions and from the Group of Seven countries. To build a custom version of the worksheet loaded with your own list of bonds, click on the Save as Template button.

You can also create a custom country risk monitor by loading a set of sovereign tickers such as {80710Z CN Equity} for Canada. Because the data is updated daily, you can see how geopolitical ratings evolve by refreshing the worksheet while headlines are crossing the Terminal. This data can provide an additional real-time signal before bond sales or holdings reports. It might help answer questions

38

INSIDE THE TERMINAL

---

Fig. 2 Go to {FLDS } to explore Seerist data for a selected ticker.

Fig. 3 To download a copy of the Seerist Geopolitical Risk Data Dictionary, run {DOCS 2232077 .

such as "Are investors shying away from longer-term bonds as civil unrest rises?" For an additional sample worksheet that comes loaded with equity tickers, go to {WSL COUNTRYRISK } for the Geopolitical Risk Monitor.

You can use the Fields function to explore Seerist data for a selected ticker. To review data for France, for example, go to {223727Z FP Equity FLDS } (FIG. 2). Set the Source to Data License.

For a copy of the Seerist Geopolitical Risk Data Dictionary, which includes detailed definitions of the different data fields, go to {DOCS 2232077 } (FIG. 3). For documentation, go to {DOCS 2196574 .

TO BE SURE, sovereign debt moves for a variety of reasons: some domestic, some international, some hedging, some speculation. Yet,

for benchmark sovereign bonds, the ability to compare bid-to-cover ratios (a metric that tracks how oversubscribed auctions are) against geopolitical risk factors can be illuminating. Demand for the benchmark US Treasury 10-year note has been steady for the past 10 years. Will that demand wane if political risk rises?

Adding formal geopolitical risk downgrades and upgrades to trading decisions can provide an additional signal about sovereign debt demand in both the primary and secondary markets. Whether you're a sovereign investor, a reserve watcher, an FX strategist or a country-risk analyst, adding this practical signal can help you gain sharper real-time insights.

Goldenberg is a market specialist for money markets, securities finance and fixed-income liquidity at Bloomberg in New York. Shumaylova is on the risk and investment analytics enterprise data team in London.

INSIDE THE TERMINAL

39

---

Macro

Track Emerging Risks by Querying Bloomberg Analyses and Data

By ALEX WISCH and CARLOS MELTZ

FOR MANY INVESTORS, 2026 started with a sense of cautious optimism. Equity markets were near all-time highs, buoyed by technology stocks. Rates hawks were beginning to believe the US economy would come through a period of higher rates without a recession. Credit market participants were flush with cash and ready to buy, causing credit spreads to tighten despite higher issuance.

After the US and Israel attacked Iran on Feb. 28, the picture changed. Oil surged, with Brent crude rising more than 60% by the end of March. Confidence about the direction of central bank policy faded as expectations of rate hikes were priced in. The equity market seemed set to dip in response to disruptions of global supply chains, but stocks have largely remained resilient.

Practitioners across asset classes are now navigating a macroeconomic regime in which traditional correlations have become unreliable. Use these tools to identify and examine areas of emerging risk and to interrogate Bloomberg's qualitative and quantitative analyses.

FIRST, FOR INSIGHT into how corporate executives are handling macro uncertainty, use the Orange Book. Like the Beige Book—the Federal Reserve's anecdotal snapshot of economic conditions across the US—the Orange Book aims to capture trends in the economy that may not be readily apparent in traditional indicators. Put together by Bloomberg Economics and Bloomberg Intelligence, it uses artificial intelligence to extract comments from earnings-call transcripts of S&P 500 companies, distilling persistent and emergent macroeconomic themes. Among them are topics such as business conditions, costs, labor markets, investment and consumer demand. Type "news on Orange Book" in the command line of a Terminal screen and hit . Click on the most recent edition.

The June 22 edition included information from 489 companies. To get answers about a sector, company or topic, click on Ask a Question in the sidebar. Then enter your question in the field that appears. For example, type "What are the technology sector's biggest risks?" and hit . The question agent will summarize insights from the document (FIG. 1).

ANOTHER NEXUS OF risk this year: the $1.8 trillion private credit market. One area of stress has been redemption requests by

investors in nontraded business development companies as AI threatens the outlook for software makers that borrowed from private lenders. You can dig into news, documents and data relating to the topic using ASKB, Bloomberg's conversational AI interface. Run {ASKB } and type, for instance, "I would like to analyze nontraded BDC investor redemptions. Please output the BDCs with the top redemptions year-to-date, their returns, margin and sector focus where applicable," then hit .

The output provides insights and citations of news, Bloomberg Intelligence proprietary data and filings (FIG. 2). Among the themes that emerge is how software exposure has been a catalyst for redemption requests. (Pro tip: When you type a general prompt in ASKB, you can enrich it by clicking the Optimize button below the query box.)

To compare technology exposure at publicly traded BDCs, use the Bloomberg Intelligence Investment Management dashboard. Type "Bloomberg Intelligence investment management" in the command line and click on the BI FLOWG match in auto-complete. Then click on BDC under Data Library. The shortcut is {BI FLOWG BDC }. Click the BDC Portfolio Exposure tab. Use the Any Sector drop-down to select Technology. To chart how selected BDCs' exposure to tech has changed over time, tick the radio button next to Chart and then tick boxes for the BDCs you want to graph.

Finally, to visualize some of the main concerns or challenges for S&P 500 companies, based on transcripts of corporate events, use the World Macroeconomic Analyzer. Type "macroeconomic analyzer" in the command line and click on the ECAN match. Type "S&P" in the field at the top of the screen and click on the AI-Powered S&P 500 Topic Trends match. The shortcut is {ECAN US DS SPX Topics }.

Use the Dimension drop-down to explore major topics including interest rates and job cuts. Generative AI, for example, saw an explosion of mentions in 2023 and remains a significant topic of conversation among tech and financial companies (FIG. 3). If you select interest rates as the topic, the analysis shows that concerns peaked in 2023. ● —With Diana Fonte

Wisch is a news application specialist, and Meltz is on the sales team, both at Bloomberg in New York.

40

INSIDE THE TERMINAL

---

Fig. 1 To dig into the Orange Book, type "news on Orange Book" and click on the latest edition.

Fig. 2 To ask a question about a topic such as redemption requests at nontraded BDCs, go to (ASKB ).

Fig. 3 To track the main concerns of S&P companies as represented by mentions in transcripts, run (ECAN US DS SPX Topics ).

INSIDE THE TERMINAL

41

---

Indexes

A Deep Dive Into Australian Equities

By ARISA MORI

AUSTRALIA'S EQUITY MARKET is highly sensitive to a narrow set of macroeconomic drivers. According to Bloomberg Intelligence, inflation explains roughly 75% to 85% of profit growth across the country's largest listed companies. Combined with commodity prices, lending growth and interest rates, these macroeconomic variables shape much of the earnings outlook for Australia's largest companies.

The structure of the market reinforces those relationships. The Bloomberg Australia Domestic 200 index, which is made up of the country's 200 largest stocks by market capitalization, allocated about 36% to financials and 25% to materials as of July 20—together accounting for more than 60% of the index. The financial sector is exposed to domestic interest rates and credit cycles, while materials is influenced by global inflation and commodity market trends.

This macro sensitivity doesn't mean the market moves in unison. Instead, it often drives sharp sector divergence. For the year through July 20, the Bloomberg Australia Domestic 200 Materials index rose 32% and the energy-sector index gained 17%, supported by stronger metals and energy prices. By contrast, the technology and healthcare indexes declined roughly 49% and 40%, respectively. The tech sector, dominated by software, has benefited little from the global artificial intelligence investment cycle.

To compare sector index returns using the Group Ranked Returns function, run {AD200P Index GRR } (FIG. 1). The sector divergence raises an important question: How representative is a broad Australian equity index of the market exposure an investor intends to capture?

Understanding Index Architecture

You can find an extensive framework for navigating the market in the Bloomberg Australia Domestic Equity index suite. To explore it, go to {IN AUNZ } (FIG. 2). The index family includes benchmarks ranging from the concentrated Australia Domestic 50 to the comprehensive Australia Domestic 600. Each has a distinct balance of market breadth, stock concentration and sector exposure.

To evaluate how these varying index characteristics impact an investment mandate, analyze them with the Portfolio & Risk Analytics (PORT) function. Load the Australia Domestic 200, for instance, by right-clicking on its name in IN and selecting PORT in the menu that appears. In PORT, you can compare portfolios and indexes side by side to evaluate sector allocations, factor

exposures and other fundamental risk characteristics. Comparing the Australia Domestic 200 with the Australia Domestic 50, for instance, reveals that the broader investment universe increases the span of market caps covered by about 25% while reducing the aggregate weight of the 10 largest constituents by 12 percentage points (FIG. 3). These architectural differences directly influence sector exposures, diversification and overall portfolio risk.

Beyond Total Return

For domestic market participants in Australia, conventional total return measures may not fully reflect their true after-tax investment experience. The reason: Unlike in many global markets, the country's dividend imputation system allows eligible investors to receive franking credits on domestic corporate dividends, which can offset the investor's taxes on the payout.

The Bloomberg Australia Domestic Equity index family thus includes grossed-up franked-return and superannuation franked-return variants, designed specifically for distinct investor profiles: tax-exempt entities and superannuation funds. Using the Comparative Returns (COMP) function, you can compare total-return and franked-return versions of the same index to isolate and better understand the contribution of dividends and franking credits to long-term investment outcomes.

For example, over the five-year period ended July 20, the cumulative return difference between the standard total return and the grossed-up franked-return variants of the Bloomberg Australia Domestic 300 index—{AD300T Index DES } and {A300GF Index DES },

NAVIGATING AUSTRALIAN EQUITIES requires looking beyond headline market returns. It demands a granular understanding of how index construction shapes market exposure, how return methodologies influence investment outcomes and how those indexes can serve increasingly specialized investment objectives. Benchmarks are no longer simply measuring the market—they can help define it. ● —With Vignesh R S, Ji Zhuang, William Lim and Jean Saw

Mori is APAC head of index product management at Bloomberg in Tokyo.

42

INSIDE THE TERMINAL

---

Fig. 1 To compare returns of sectors of the Bloomberg Australia Domestic 200 index, run (AD200P Index GRR ).

Fig. 2 To explore the Bloomberg Domestic Equity index suite, go to (IN AUNZ ).

Fig. 3 To analyze how an index's characteristics compare to a portfolio or benchmark, go to (PORT ).

INSIDE THE TERMINAL

43

---

Paradoxes

Want to Pick the Better Trader? Here's One Statistical Trap To Look Out For

By FRANCESCO TONIN

THERE'S AN OLD JOKE: The head trader on a London trading floor was given the task of selecting candidates for a position from a huge stack of résumés. He promptly split the pile in two and threw half in the trash. As colleagues stared aghast, he explained, "I don't want to hire somebody who's not lucky!"

For a recruiter trying to figure out what type of person a candidate is, a résumé can be a bit like an encrypted message. Based on a list of stylized accomplishments, you try to decipher the applicant. Yet a statistical paradox with connections to the world of cryptography adds a further wrinkle: Even comparisons of candidates' track records can yield contradictory interpretations, according to what's known as Simpson's paradox.

EDWARD SIMPSON WAS an English statistician who during World War II worked with Alan Turing at Bletchley Park, the signals intelligence operation famed for breaking the German Enigma code. After the war, Simpson laid out in a 1951 paper "The Interpretation of Interaction in Contingency Tables" how statistical associations can change depending on how data is grouped.

How does Simpson's paradox work? Consider a table showing the performance of two traders on days the market rose and on days it fell (FIG. 1). Their track records don't cover the same days, so it's not a direct comparison. Imagine then that you have to pick one of them to fill in on your team tomorrow. You don't know whether the market will be up or down tomorrow, but you have this historical data.

Trader A has been in the market on 20 days when the market was up. She beat the market on 18 of those days, for a hit rate of 90%. For both up and down days, Trader A's hit rate is higher than Trader B's. So you deduce then that your better choice is to pick Trader A. That is, until you look at the total numbers and notice that Trader B has a better hit rate there. How can this be?

The incongruity seems to emerge from looking at the data from two different perspectives. First: If you know it will be an up day, then Trader A would be your choice because of her higher hit rate. But if you know it will be a down day, then Trader A again would be your choice. You feel like the knowledge of market direction should therefore make no difference at all in your decision.

Second: If, however, you don't know how the market will perform, you should look at the aggregate numbers, according to which Trader B should be your choice.

To resolve the seeming contradiction, notice that in the first perspective, you look only at hit rates, but not at the actual numbers of days. Trader B has had a longer track record, and that's a fundamental ingredient for Trader B's total hit rate. In effect, the two ways of looking at it make use of different information.

For some visual intuition about what's going on here, consider a chart. Plotting days beating the market on the Y-axis and total days on the X-axis, you can then draw a vector for Trader A on up days. The slope of the vector represents the hit rate. Do the same

Fig. 1

44

INSIDE THE TERMINAL

---

Fig. 2 A chart that plots outperforming days against total days provides visual intuition.

Fig. 3 When you sum the up and down vectors for the two traders, however, the picture changes.

for Trader B on up days, and it's apparent that the incline of this vector is not as steep—though it's significantly longer: Trader B has been in the market on many more up days. Repeat with vectors for the down days (FIG. 2). Similar comparison. Yet when you add each of the traders' two vectors, it becomes clear that Trader B's total slope is steeper; her overall hit rate is higher (FIG. 3).

In fact, if you assume that Trader A will keep her 90% hit rate on up days, you can extend the vector and discover how much time it would take for the two traders' total hit rate to become equivalent—when the sum of Trader A's vectors has the same slope as the sum of Trader B's.

For an animation illustrating this effect, which was vibecoded using BQuant Desktop, Bloomberg's sandboxed Python environment, reach out to the author at ftonin5@bloomberg.net.

CUTTING-EDGE PRESENTATIONS on topics related to financial math can be found monthly at the Bloomberg Quant Seminar. Chaired by Bruno Dupire, Bloomberg's head of quant research, and held at Bloomberg's New York headquarters, BBQ is the world's largest regular event of its type.

You can sign up to get invitations at professional.bloomberg.com/explore/quant-seminar-series/.

To watch past seminars on the Terminal, run (SMNR ), type "BBQ" in the Search field and click the Exact Match. When upcoming seminars are posted here, you can also sign up for them by clicking on the Register button. ●

Tonin is the product manager for FX options electronic trading at Bloomberg in New York.

INSIDE THE TERMINAL

45

---

Features

Bloomberg Markets

PHOTOGRAPH BY BENEDICT EVANS

46

---

Cans roll off an assembly line in Pennsylvania

Trump and the Tin Can	p 48
Al Rewires South Korea	p 58
SaaSpocalypse Now?	p 64
Courting Stock-Shy Europe	p 70

Economics

---

By SHAWN DONNAN PHOTOGRAPHS BY BENEDICT EVANS

---

WHAT HAPPENS WHEN TARIFFS COME TO THE HEARTLAND

---

In the chaotic history of Donald Trump's tariffs,

Commerce Secretary Wilbur Ross' appearance on CNBC on the morning of March 2, 2018, is a mostly forgotten footnote. The president's surprise announcement the day before that he was planning a 25% tariff on imported steel had sent markets tumbling. Ross, a Wall Street titan, was there to try to convince nervous investors that their fears were overblown.

Four minutes in, in a gray suit, his tie slightly askew, the billionaire reached for a prop on the desk before him. "What I would like to do, though, is to emphasize again the limited impact," Ross offered in his deliberate monotone, raising his left arm and cocking his wrist. "This is a can of Campbell's soup."

Ross said he'd gone out that morning and bought a can of chicken noodle at a Florida 7-Eleven for $1.99. By his calculation, it contained just 2.6¢ worth of steel. "If that goes up by 25%, that's about six-tenths of one cent on the price of a can of Campbell's soup," he went on. "Who in the world is going to be too bothered by six-tenths of a cent?"

The small costs would come with immense benefits, Ross argued. "You're talking tens and tens of thousands of jobs being created, and hundreds and hundreds of millions of dollars of capital investment coming," he declared.

The levy on steel marked the first major tariff volley of a presidency that, in the economic realm at least, has been defined by import taxes. In late July, Trump slapped duties as high as 12.5% on imports from 59 countries and the European Union after the US Supreme Court earlier rejected the legal rationale for much of his protectionist regime. Trump's theory remains that tariffs—taxes paid by

Ross, then commerce secretary, illustrating his point on CNBC in 2018

importers on goods purchased from abroad—will protect US industry and drive companies to invest in domestic factories. The goal is to spur an American reindustrialization and an economic rebirth in parts of the country that spent decades losing manufacturing jobs and people. The idea helped carry Trump to the White House in 2016, and he's doubled down on it since returning last year.

Eight years into the experiment, though, the story hasn't turned out quite as Ross, or Trump, foretold. What has actually happened to the tin can, how it's made and the future it faces, says a lot about how the president and his often unpredictable use of tariffs have shaped the US economy since his first election. It's a story that's not just about the economics of tariffs but also about unintended consequences and unfulfilled promises.

Costly Cans

Change in US consumer price index since March 2018

✓ All items ✓ Food at home ✓ Canned fruits and vegetables

Figures based on seasonally adjusted indexes Source: US Bureau of Labor Statistics

TALL AND WIRY, Robert Gatz moves with the purpose of an impatient manager. To follow him on a tour through the production floor at the Can Corporation of America's main, 350,000-square-foot plant is to go speed-walking with industrial cacophony as the soundtrack. He weaves through the floor and in and out of safety gates as he describes machines and plucks the odd can from the line to illustrate a precise robotic weld or the wrinkles that add structure and strength.

Each year, Can Corp. and its 350 employees turn out just shy of a billion tin cans in 200 different sizes that are destined to be filled with everything from coffee to industrial adhesives. "Give your product a touch of class by packing it in an attractive and safe metal package," goes the sales pitch on the company's website. Can Corp.'s founding Giorgi family still owns the business, based near Allentown, Pennsylvania, which was born in 1976 out of frustration with the quality of cans the family was buying for its mushroom business.

The case for the tin can has remained much unchanged since Napoleon put out a call for a new way to store and ►

50

BLOOMBERG MARKETS

---

Can Corp.'s Gatz

---

preserve food for his French army in the late 18th century. Seal food away from the elements, and it won't go bad. Do it in a rigid metal container, and you can stack it high in warehouses for years at a time until your army needs it.

Gatz is biased, of course. He's a tin can man. The Can Corp. vice president and general manager argues the receptacle has no real rivals when it comes to storing food. "I always tell everybody that you can go home tonight, go into your pantry and take a look at some cans that you have in there. You probably bought a can of beans or a can of soup or whatever. And its use-by date, or eat-by date, is 2023. It's fine. Eat it. There's nothing wrong with it."

There's an economic logic to making tin cans in the US. Shipping empty cans means shipping air, Gatz says, and subjecting the metal containers to potential corrosion.

So manufacturers serve markets close to them and rotate production with the seasons, catering to the harvesting of fresh fruit and vegetables in summer and the canning of soups and other prepared foods in winter. Trump's steel tariffs, which he hiked to 50% after he returned to office in 2025, have disrupted those economics over the past eight years. US tin can production costs have spiked since the first steel duties went into place in 2018.

Cans are made from tinplate steel, produced by rolling steel and thin layers of tin together into sheets. And while the US steel industry has been investing in other areas, it's scaled back production of tinplate, which represents a tiny and low-margin segment of the market and employs only a few thousand people. Canmakers now rely more on pricier imported tinplate.

52

---

As a result, Can Corp. has raised prices by double-digit percentages this year, leading to “a lot of difficult conversations” with customers, Gatz says. The price of an empty can produced in the US is up almost 80% since March 2018, while consumers pay almost 50% more for canned fruits and vegetables, official data shows. Over the same time, the overall price of food has risen far less, suggesting tariffs are at least partly responsible for the higher prices.

Facing higher costs for the metal as a result of Trump’s tariffs, Can Corp. and other manufacturers have tweaked designs to use thinner steel. Gatz and other industry executives also spent much of 2025 and early 2026 making frequent visits to Washington, pleading the case for tariff relief with the Trump administration and people in Congress. The result: a lot of sympathy but no relief. Gatz isn’t getting his hopes

Abbett (far left) with sons Casey and Austin on their Indiana farm

Hard Times

Employment in iron and steel mills and ferroalloy production in the US

Source: US Bureau of Labor Statistics industry productivity data

up that the tariffs will go away. “It’s happened,” he says, in a nod to Apple Inc. and other companies that have secured exclusions. “But it’s tough to get that to happen.”

So, by the summer, America’s canmakers decided to try a different tactic. Rather than ask for the removal of tariffs on imported steel, they began focusing on placing a tariff on a competitor that had avoided the levies: the producers of already-filled cans made overseas, which were packing the shelves of domestic supermarkets.

Their shift in approach recognized the reality of tariff politics in the Trump era. It’s easier to get new tariffs than to chase away existing ones. “We think that it’s more palatable to the administration to address foreign competition than it is to lower the tinplate tariff,” says Scott Breen, who leads the Can Manufacturers Institute, the industry’s lobbying group in Washington.

IF THERE’S a family farm version of the American dream, Glenn Abbett is convinced he’s living it. He works on land his father started cultivating in the 1960s. Two of his three adult sons have homes nearby and work alongside him. Tin cans help keep Abbett’s dream alive, no small feat given the rising costs and falling commodity prices battering many family farms.

Since 1982, Abbett Farms in La Crosse, Indiana, a 90-minute drive from Chicago, has planted hundreds of acres of tomatoes each spring. All are under contract to Red Gold Inc., which each summer fires up three canneries and hires about 600 seasonal workers for a 24-hour-a-day, 60-day surge to can more than 400,000 tons of fresh tomatoes. “It’s been by far the most profitable venture that we are involved with on the farm,” Abbett says. “It’s been kind of a godsend.”

Because they’re grown under contract, tomatoes are a predictable income stream that occupy less than 10% of the farm’s 8,000 acres yet yield more than 40% of its annual revenue. They also insulate Abbett’s business from wild swings in commodity prices caused by speculators. “With tomatoes, that’s just taken out of the mix completely,” he ►

THE BALLAD OF THE AMERICAN TIN CAN

53

---

says. Partly because of the rising cost of tin cans, however, Red Gold contracted Abbett Farms to plant just 465 acres of tomatoes this summer, 10 fewer than it did last year and far below Abbett's ideal of 600.

For Abbett it's been a real-world education in how tariffs work. He voted for Trump and supports his efforts to bring manufacturing back to the US. But he didn't foresee his business bearing the cost and believes it's an unintended consequence of Trump's policy. "I want a strong economy, because I believe the strength of our economy ultimately is the backbone that holds together our ability to defend ourselves in the future," Abbett says. But "this tariff is obviously bothersome. It's like he hasn't heard from the right person. Or he hasn't heard the voices loud enough."

Because of tariffs, Red Gold is paying as much as 17% more for its cans this year depending on the size or manufacturer, says Gary Petersen, senior vice president for sales. It's also seeing a threat to revenue. Big institutional customers are turning to cheaper imported canned tomatoes from Italy and Egypt, he says.

Red Gold is a fourth-generation family-owned company that has relationships with farmers going back four or five generations as well, Petersen says. So it's not slashing how many tomatoes it buys just yet. Tomatoes that aren't canned can be turned into paste, ketchup, sauce and other products, he says. Still, he could foresee cutting jobs,

as well as tomato purchases. So the company is lobbying alongside other food processors for tariffs on imported canned tomatoes.

Then again, there's only so much tariffs can do. In the spring the can industry failed to get steel tariffs expanded. At the same time, the Trump administration was easing duties on food imports as it tried to address voter concerns about grocery prices before November's midterm elections.

Imports of canned peaches have been subject to a 17% tariff for decades, and, since the first Trump administration, a 25% additional duty has applied to Chinese peaches as well. And yet "we are still seeing significant volumes of imported products entering this country," says Rich Hudgins, chief executive officer of the California Canning Peach Association. In fact, he says, "the steel tariffs just have kind of a counter-productive effect of compounding the cost advantage that our competition in China and our competition in Greece deal with today."

WHEN YOU DRIVE into Weirton, West Virginia, it doesn't take long to run into a sleeping giant. Weirton's steel mill employed more than 14,000 people at its height in the 1970s and was one of the largest tinplate producers in the world. It retains a dominant physical presence in the town. Its buildings, painted a vivid blue, line its streets. But its current

Part of Weirton's now-closed steel mill, with the new Form Energy battery plant (left)

---

Baker in Weirton

owner, steel giant Cleveland-Cliffs Inc., which acquired the mill in 2020, shut it for good in early 2024, laying off more than 900 employees. These days the facility is an emblem of the declining US production of tin mill steel that has materialized alongside Trump's tariffs. "We did everything we could possibly do to try to keep it afloat. And it just wasn't enough," says Ron Baker, who first went to work at the mill in 1972 and now leads the 25-Year Club, which brings together former employees for a picnic each July. "We wanted it to be there for years, for generations. And it just wasn't in the cards."

Trump's tariffs just haven't generated the promised steel industry job boom. US mills employed 85,400 people in 2025, only 1,300 more than in 2018. The day Ross went on TV that year, a dozen mills in the US produced the tinplate that goes into cans, and domestic can manufacturers imported half the steel they used. Today there are just three active tin mills, and 80% of the steel used in cans is imported.

Cleveland-Cliffs shut the Weirton mill in 2024 after the US International Trade Commission, the government body that adjudicates American companies' allegations of unfair competition from imports, rejected the company's bid to clamp down further on foreign competition. Trump's

first-term tariffs, which Joe Biden's administration extended, weren't enough to justify keeping the plant open, the company said at the time. It needed more duties to compete. "I want to make this point as clearly as possible: The success or failure of this trade case will determine whether the United States will continue to produce tin mill steel," CEO Lourenco Goncalves testified to the USITC in January 2024.

Mark Glyptis, who's served more than 30 years as president of Local 2911 of the United Steelworkers, represents the Weirton mill's employees. He saw the USITC decision as an existential blow. Decades of trade battle memorabilia, including a red-white-and-blue "Free Traders Are Traitors" protest poster, plasters his office. If the USITC had imposed additional tariffs, the mill might still be operating today, he says. Even the 50% tariff now in place would have helped.

The USITC ruled against Cleveland-Cliffs in part because the aging Weirton mill produced tinplate that was too narrow for the contemporary needs of can producers. Glyptis lobbied the owners as far back as the 1980s to invest in the mill so it could produce a wider product. It "would have given us a very good chance of survival," he says.

Glyptis campaigned for Trump in 2024. "I thought ►

55

---

United Steelworkers' Glyptis in his Weirton office

Trump would help us," he says. But now he feels let down. The president's volatile trade policy has made it hard to find a buyer for the mill, which will cost at least $300 million to revive. The city is moving on. For more than a century, steel and tin helped define Weirton's identity. City officials say they now have high hopes for the bright white factory that opened in 2024 across from the idled tin mill. There, startup battery maker Form Energy is employing 400. "Steel is what got us here," says Dean Harris, Weirton's mayor and a 48-year veteran of the steel mill. "But it's not our future."

TINPLATE WAS ONCE considered a strategic priority for the US. Alexander Hamilton listed its production among the essential industries the country needed to develop when, as the first Treasury secretary, he sent his 1791 Report on the Subject of Manufactures to Congress. When Ernest Weir opened the Weirton mill in the town that eventually took his name, in 1909, the US market for tin mill steel went far beyond cans. American homes had tin ceilings, and the families that lived in them ate off tinware plates with cutlery made out of tinplate.

These days tin mill steel is used for food containers, aerosol cans, oil filters and electronic components but not much else. Tinplate accounts for about 1% of the US steel market, which relies much more heavily on demand from the auto industry and construction than canmakers.

Jason Miller, an economist and supply chain expert at Michigan State University, says it makes little sense to put tariffs on the steel used to make tin cans. "It's not a product that US steelmakers really want to produce," he says. "So you're not even providing tariff protection for a product they really want to make. And you're not protecting many jobs in steelmaking." Moreover, tinplate users employ many more people. About 78,000 people work in US canneries, almost as many as in all of the country's steel mills, Miller points out.

Scott Paul, president of the pro-tariff Alliance for American Manufacturing, an advocacy group with links to the steel industry, says Trump's tariffs have helped stabilize the industry. He'd like to see even higher levies on tinplate. In his view they'd protect both steel companies and communities like Weirton that depend on mills.

Paul points to a new tinplate test case. In April, US

56

BLOOMBERG MARKETS

---

Steel Corp., owned by Japan's Nippon Steel Corp., asked the USITC for new tariffs on tinplate imports from China, Taiwan and Turkey. The company also said it would invest as much as $20 million to restart an idled tin mill in Gary, Indiana. Still, Can Corp.'s Gatz says, even that revival will leave domestic canmakers mostly dependent on imported steel, because the amount of tinplate Gary can produce "is minuscule compared with the overall problem."

THAT CAN OF CAMPBELL'S chicken noodle soup Wilbur Ross held up on TV would cost you $3.49 at 7-Eleven today, 75% more than he paid in 2018. In a statement, Campbell's Co. said it doesn't set the retail price for its cans and has been working to "help keep our products affordable for American consumers" in response to tariffs and inflation. Kush Desai, a White House spokesman, blames the Biden administration for exempting too many companies and countries from tariffs and failing to address affordability: "The story here really is about how Biden's inflation crisis increased food prices."

Today, Ross isn't convinced that the American steel industry's future is in tin cans. Asked why things turned out as they did, he points to the rising cost of tin and the surging price of the fruits and vegetables that fill those cans. Making tinplate is also finicky, offering only a thin profit margin for producers, he says.

He speaks from experience. For a few years in the early 2000s, Ross and his International Steel Group owned the Weirton mill, until he sold his steel company to Indian billionaire Lakshmi Mittal for $4 billion in cash and stock in 2005. ISG struggled to make the Weirton mill profitable and cut costs, Ross says. "The acquisition we made of Weirton turned out not to have been our best idea."

Still, Ross has taken away a broader lesson about tin cans from what's happened since he went on TV that day and brandished his can. "One thing it proves is that tariffs don't necessarily solve every problem the world has," he says. "And that, I think, will always be true."

Donnan, a senior reporter based in Washington, covers economics.

A Can Corp. worker on the production line

---

THE LAND OF AI

By HEESU LEE, SOO-HYANG CHOI, YOOLIM LEE and SOHEE KIM

ILLUSTRATIONS BY DANI CHOI

---

HOW THE TECHNOLOGY IS REWIRING SOUTH KOREA

---

Two wall-size maps in Dongtan, South Korea, render every street and apartment block in meticulous detail.

Between them stands real estate agent Mark Yoon, laser pointer in hand. Clients drift in and out of his shop-front office asking about apartment prices, commuting times and school districts in the city, about 30 miles south of the country's capital of Seoul. Yoon answers by tracing laser lines across the maps. Eight years of selling homes here have taught him that almost every conversation eventually arrives at the same destination: semiconductors.

The reason is visible on Yoon's maps and just beyond his office windows. Nearby shuttle buses ferry engineers to Samsung Electronics Co.'s vast semiconductor campuses in Hwaseong and Giheung, and highways put rival SK Hynix Inc.'s operation within reach. That makes Dongtan a residential crossroads for South Korea's chip industry, which pumps out about 80% of the world's high-bandwidth memory chips that are powering the artificial intelligence revolution.

Yoon recently brokered one of Dongtan's priciest deals: an 84-square-meter (904-square-foot) apartment that sold for more than 2.2 billion won ($1.5 million). People in Dongtan can afford expensive property because of news that drew worldwide attention. Some rank-and-file SK Hynix and Samsung chip workers are expected to collect bonuses of at least $400,000 for 2026. Even before the money reached employees' bank accounts, asking prices for apartments near Dongtan station climbed by hundreds of millions of won in a matter of weeks.

"We used to suspect Dongtan's housing market moved with the semiconductor industry," Yoon says. "After this year's rally, I think it's been proven. The two are 100% linked." And so, increasingly, are Korea's economic fortunes.

Around the world, the rise of AI has inspired skepticism,

wonder and dread. But it's hit Korea differently. In the US, an elite circle of Stanford University and Ivy League graduates is ushering in the era from Bay Area office suites. In Korea, the boom radiates from factory floors.

The success of Korea's chip industry is astonishing. SK Hynix's shares surged more than 1,000% in the 12 months leading to their June 22 peak, while Samsung jumped some 500%, propelling the value of stocks traded on the Korea Exchange's main index past markets in Canada, Germany, the UK, France and India earlier this year. In July, SK Hynix made a blockbuster Nasdaq debut, two months after its market value hit $1 trillion. AI marks the latest success of Korea's flavor of industrial policy, in which governments nurture critical businesses. A succession of administrations used the approach after the war that divided the country from the North in the early 1950s and left it poorer than many sub-Saharan nations. South Korea backed steel, shipbuilding, automobiles, consumer electronics and semiconductors, helping make the Asian nation of 51.6 million wealthier than the average member of the European Union.

In late June, South Korean President Lee Jae Myung staked his legacy on a plan to transform the nation's less developed southwest into a new memory hub while securing leadership in data centers and robotics. Flanked by the bosses of Samsung and SK Hynix, his ministers announced semiconductor investments totaling at least 1,350 trillion won. Lee's administration is also pushing for the country to create homegrown AI foundation models to compete with the likes of Anthropic, DeepSeek and OpenAI, and ultimately help make South Korea an alternative to the US and China as they dominate the global AI race. Searching for the best model, the government is backing a tournament nicknamed "AI Squid Game" after Netflix's award-winning dystopian TV series. The contestants face evaluation and elimination every six months.

AI has begun to reorder South Korean society, changing expectations about fairness, careers and even culture and dating. Yoonsoo Lee, a Seoul National University economics professor, says the technology is likely to inspire debate over the need to redistribute wealth, comparing it to upheavals after the advent of railways, electricity and computers. "AI is unlikely to benefit everyone equally, and there will be clear winners and losers," Lee says.

Yoon in Dongtan

PHOTOGRAPH BY TINA HSU FOR BLOOMBERG MARKETS

60

BLOOMBERG MARKETS

---

Bumki Son, an economist at Barclays Plc, says it's still too early to tell how lasting the technology's impact will be: "The key question is whether demand will remain structurally stronger or eventually revert to past patterns—or even fall back into another downturn."

This bounty has sparked concerns as well as celebration. Proposed new chip factories in the drought-prone southwest region of Honam have raised questions over water supplies. The Korea Federation for Environmental Movement says water allocation should balance industrial development with agricultural and ecological needs, rather than prioritize corporate demand. In June, President Lee said the government would move forward only with a thorough assessment.

The more than 98% of workers who aren't in the chip sector are increasingly feeling like bystanders to the bonanza. In a bid to spread the gains, the government plans to establish a dedicated fund to finance long-term investments in strategic industries, education, regional development and youth.

The central bank has warned that gains from technology, equity markets and AI are disproportionately benefiting the affluent. Regulators have also moved to rein in speculative trading by small investors, who've been drawn in by spectacular gains in AI-linked stocks. They've bet heavily, leading to sharp market swings, including big losses since the market's June 22 peak.

President Lee's top policy adviser, Kim Yong-beom, has floated the idea of a "citizen dividend" that would use

excess tax revenue generated by the AI boom to invest in young people and future strategic industries. He posted on Facebook in May: "A nation may become wealthier, but wealth does not automatically spread across society. The central question in the AI era is no longer growth itself, but how to socially stabilize the extraordinary profits it generates."

ON A FRIDAY AFTERNOON in late June, shoppers wandered between Saint Laurent handbags and display cases of Omega watches at Lotte Department Store's Dongtan branch. Upstairs, VIP lounges hummed as customers lingered over afternoon tea. Many work in semiconductors. Applications for Samsung employee-only credit cards, which come with discounts at the Dongtan store, have been increasing by about 100 a month. More than 8,000 have been issued since 2021, says Hyunwoo Kim, who leads the branch's sales planning team. "Our VIP customers are noticeably younger here," Kim says. "The average age is in the mid-40s."

That's unusual for a department store, where the biggest spenders are typically older. Sales rose 25% in the first six months of 2026 from a year earlier, and luxury goods climbed 40%, among the fastest sales growth of Lotte's stores nationwide. Kim expects the momentum to continue as Dongtan's population grows and its commercial district matures.

Cholmin Kwon, a taxi driver who spends his days ferrying Samsung employees between apartment complexes, ►

THE LAND OF AI

61

---

train stations and the sprawling chip campuses, says bonus season has changed the tone of the conversations in his back seat. “I’ve been hearing Samsung employees talk about their bonuses all the time,” he says. “Most of the conversations are about buying a home or moving into a better apartment.”

Five years ago, Kwon remembers people questioning why a high-end department store such as Lotte had opened in what many still regarded as an ordinary commuter town. Today, he sees restaurants filling up, more expensive places drawing customers and entrepreneurs from other cities arriving to scout locations for new businesses. “It feels like more businesses are coming because people expect Dongtan’s economy to keep growing with the semiconductor industry.”

As the cabbie attests, the AI boom is creating a new social elite. For 15 years, Eunsun Kang has watched the dating market elevate one profession after another—civil servants, teachers. Doctors and lawyers have long occupied the upper ranks. Each cycle, she says, reflects what young Koreans believe offers the safest path to building a life together. Kang, a senior manager at Gayeon, one of the nation’s largest matchmaking firms, says chip employers are having their moment. “Right now is the peak,” she says. “I’ve never seen semiconductor workers this popular in my 15 years doing this.”

Women who once insisted on meeting only doctors or lawyers are now actively asking to meet engineers from SK Hynix and Samsung, she says. At the same time, many of

those engineers have become more selective, aware that their standing in the dating market has improved. “A few years ago, some of these introductions wouldn’t even have happened,” she says. “Today, if someone works at a chipmaker, meetings that used to be difficult are now possible.”

The new hierarchy has become fodder for popular culture. In a recent episode of Saturday Night Live Korea, a salesperson at a luxury boutique dismisses a poorly dressed customer, until the man unzips his jacket to reveal an SK Hynix company vest underneath. The clerk immediately changes her tone, greeting him as “Lord Hynix.” The joke has caught on. A seller on secondhand platform Danggeun Market recently listed what appeared to be an SK Hynix company jacket as “the ultimate blind-date outfit.” The listing quickly spread across social media.

That SK Hynix has become so totemic is surprising, given that 15 years ago, the long-ailing company then called Hynix Semiconductor had twice failed to find a buyer until conglomerate SK Group submitted its bid at 4:53 p.m. on Nov. 10, 2011, only seven minutes before the deadline, according to Super Momentum, a book that chronicles SK Hynix’s rise from underdog to AI leader.

In July, at the company’s Nasdaq debut, SK Group Chairman Chey Tae-won recalled that SK Hynix was widely regarded as a dangerous asset: a capital-intensive manufacturer trapped in a notoriously volatile boom-and-bust industry.

Now, the appeal of SK Hynix and its peers is reaching

62

BLOOMBERG MARKETS

---

even into classrooms. For decades, Korea's brightest students pursued medicine. In the 2026 admissions cycle, it was harder to get into five semiconductor programs backed by Samsung and SK Hynix than those in the natural sciences at Seoul National University, the country's equivalent of Harvard University, according to Jongro Academy, a private college-preparatory institute that tracks publicly available admissions data.

Seunghyun Jeong, 25, is betting his future on chips. After graduating with an electrical engineering degree, he enrolled in a semiconductor training program jointly run by KAIST, a Korean sciences and technology university, and Hwaseong City. Jeong says he's more confident about his prospects as major chipmakers expand hiring than he was during the industry downturn three years ago. He doesn't hesitate to say where he'd want to work. "I'd prefer SK Hynix," he says. "From what we've heard, the performance bonuses are better." Yet even as he tries to ride the AI wave, Jeong is mindful. "I am optimistic about the industry," he says. "But we still have to think about what happens when this boom eventually cools."

AS IN MUCH OF THE WORLD, AI is roiling Korea's cultural industries. The worldwide popularity of K-pop bands such as BTS and the survival show Squid Game have been an economic boon before the meteoric rise of large language models. Music, TV, films, video games and other cultural exports generated 161 trillion won in revenue and more than $15 billion in exports last year. They employ about 690,000 workers, more than double the combined number at Samsung and SK Hynix.

South Korea's entertainment companies are building virtual K-pop idols and AI-generated music videos, as well as digital influencers who can perform, livestream and interact with fans around the clock. Galaxy Corp., the studio behind K-pop legend BigBang's leader, G-Dragon, says it views AI not as a replacement for human stars but as a way to create new intellectual property, expand global reach and develop lower-cost entertainment franchises that can generate revenue alongside traditional artists.

In Hollywood, many writers and actors are resisting AI. In Korea, the industry is embracing the technology as it experiences a movie slump that's among the world's steepest. With

Oh Tae-hee in Seoul

A Wild Ride

Kospi (South Korean stock market index)

Level

Member weight

Sources: (KOSPI GP ), (KOSPI MEMB )

Korean box-office revenue down more than 40% since the pandemic, filmmakers are increasingly cutting costs by using AI (including Kling AI, Runway AI and Google's Gemini) to generate storyboards, visual effects and entire scenes.

In July, the city of Bucheon—on Seoul's western fringe, and former home of the nation's first semiconductor plant—hosted the 30th annual Bucheon International Fantastic Film Festival. Organizers screened AI-generated films, led industry summits and hosted an AI filmmaking academy with Seoul Broadcasting System and workshops in schools. The goal: training 10,000 AI creators by 2029.

Last year a short film by Oh Tae-hee caught the festival's attention. With no formal training or industry connections, he'd taught himself how to use AI to make movies while still working his day job as a pharmaceutical marketer. This year he returned to the festival as a juror. "People learn by experimenting, and creators share techniques with each other," he says. "That's how the ecosystem is forming." He says he's spending about 2 million won a month to subscribe to about a dozen AI video generation platforms and earning tens of millions of won per project.

Korea's film industry crossed a milestone this summer with the commercial release of its first AI-generated feature films, and more AI and hybrid productions are scheduled to premiere in theaters later this year. "Just as 'silent,' 'color' and 'digital' eventually disappeared from how we describe movies, AI will also be dropped," says Cho Yang-il, the festival's AI adviser. "In the end, there will only be films. Everyone believes AI filmmaking can become a business. They're just waiting for the breakthrough that proves it."

Yet, as in the US, some, including Park Chan-wook, remain uneasy. The Korean filmmaker is known for No Other Choice, a pitch-black satire about corporate restructuring, and the HBO series The Sympathizer, based on the Pulitzer Prize-winning espionage novel.

At Korea's Busan International Film Festival last year, he said he hoped AI will remain "an extension of our toolbox." But he warned it "could also take away many jobs and fundamentally alter the aesthetics of cinema—and that fills me with fear."

Heesu Lee, Choi, Yoolim Lee and Kim report from Bloomberg's Seoul bureau.

THE LAND OF AI

63

---

The Software Credit Crunch

64

---

By PAULA SELIGSON and MICHELLE CHENG

Will an

irresistible disruption meet immovable debt?

ON WALL STREET, there's a disaster scenario known as the "SaaSpocalypse" that goes something like this.

The financial industry—and in particular the growing titans of lever- aged buyouts—made a generational bet on software companies. A reliable stream of subscription revenue from software-as-a-service, or SaaS, pro- vided cover for aggressive investments in the era of rock-bottom interest rates. Software became one of the most popular subjects for private equity acquisitions in years marked by fren- zied dealmaking, while a new class of lenders arose to provide easy credit to fuel the purchases. Everything would be fine, as long as the business model remained intact.

With the rise of artificial intelli- gence, that model and the financial edifice built on top of it are being threatened, with consequences that could extend beyond the software com- panies themselves. Rapidly developing AI tools, investors fear, could replace many existing tech products. This risk affects all kinds of software companies, but the ones owned by private equity buyout funds are facing piles of debt and market headwinds.

Buyouts have always been risky—what's different now is the worry that a wide swath of wagers could go sideways at once. "The senti- ment overhang to software is difficult to disprove in this environment because we don't know what the pace of the technological advancements will be," says Amanda Lynam, chief credit strat- egist at Goldman Sachs Research.

A Match Made on Wall Street

Software companies were particularly attractive to buyout funds because they had relatively low costs—and high profit margins—while steady subscription revenue allowed them to make interest payments if they bor- rowed heavily. This debt is a crucial part of the private equity business model.

In a leveraged buyout, a private equity fund puts up some of the pur- chase price for a business and borrows for the rest, though the debt goes on the target company's balance sheet. That structure can magnify returns for the buyout fund if the company's

value rises. But it can also leave the acquired company more fragile, because it has to pay back all that bor- rowing, plus interest.

Traditionally for large buyouts, dealmakers would tap the junk-bond and leveraged loan markets, where banks arrange deals and then sell the debt to investors. But in recent years, a more opaque private credit market became an important source of funding. The money came directly from investment funds that raised cash to make loans. The setup gave private equity firms a new group of lenders competing for their business.

Private equity and credit invest- ments in software are now weighing on some of the biggest asset managers that led the charge and have ended up in the portfolios of insurers and pensions, which are major investors in private credit funds. Retail investors also began snapping up new loan funds marketed by wealth advisers.

Junk bonds and leveraged loans are traded, providing an outside market measure of their worth. But private equity firms are slow to mark down the value of their equity invest- ments, and even when they do, it's usually only disclosed to their inves- tors. Private credit loans don't typically trade, and while some funds do disclose marks on the value of the debt, these are widely viewed as lagging and poten- tially too optimistic.

This has resulted in a sort of dark matter floating around the financial system, leaving investors uncertain ▶

PE's Software Bet

Software's share of the value of US private equity deals

*As of June 30. Source: PitchBook

65

---

Buying a $6 Billion Company for $3 Billion

How a leveraged buyout works

Source: Bloomberg

about the size and the location of the overall risk. Lee Robinson, one hedge fund manager famous for spotting risks during the 2008 financial crisis, is betting against insurers because of the potential hit their private credit holdings could take. Even banks have ended up with some exposure through lending to the credit funds themselves.

AI won't hurt all software companies. Customers may not be able to walk away from technology they've worked with for years. Many software companies could win out by integrating AI into their products. Tech-oriented investment firm Thoma Bravo, for example, struck a strategic partnership with Alphabet Inc.'s Google Cloud to help its portfolio companies accelerate their adoption of AI. Even if some businesses end up becoming obsolete, a private equity fund only needs a few big winners to make up for a dud. And most borrowers are still growing revenue and keeping up with payments.

But the AI shift is just starting, and investors don't know how to predict winners and losers. It's also a

precarious moment to be leveraged because rates have risen. Buyout funds often use loans whose costs rise and fall, or float, based on interest rates. Meanwhile, fears that software companies will be disrupted by AI have led lenders to demand higher yields to refinance the debt as it comes due.

Many software-sector loans were taken out based on the assumption that the companies would be able to expand their subscriber base. Even if the software companies don't collapse, a decline in revenue (or even a slowdown in its growth) could be enough to put the debt under pressure.

"If you strip away the capital structure and the AI disruption risk, the operating profile of most of these companies looks really healthy," says Alen Lin, a senior director at Fitch Ratings covering North America technology credit. "It is that high financial leverage, the high interest expense, that's dragging down the overall financial metrics." Fitch found that half of software companies it rates are at low risk of AI disruption, while 9% are at high risk. Meanwhile, anxious debt

investors have sold their holdings because of the threat posed by AI, which has been visible in the value of leveraged loans (first chart on page 68).

AI risk is also haunting public software companies such as Salesforce Inc. and Workday Inc., and their stocks took a hit during the SaaSpocalypse selloff. Although these listed companies are generally larger than their PE-owned peers, they don't carry the same high levels of debt, relative to their size, giving them more room to maneuver around any AI disruption.

If things start to go wrong, it could set off a chain reaction of layoffs in office parks and skyscrapers across the US as software companies slash costs. And if untenable debt loads lead to bankruptcy or restructuring, losses would also hit all the exposed investors and lenders. Private equity firms could lose their whole investment, while big institutions and retail investors who've put money into the loans would find them impaired. That might lead lenders to pull back in ways that could have a broader chilling effect on the economy.

"There is definitely a 'returning

66

BLOOMBERG MARKETS

---

"There is definitely a 'returning to Earth' situation that has to happen here in the coming 18 months to two years"

A Shrinking Cushion

How debt becomes a problem

① Imagine a company known for a software tool for which it sells per-user subscriptions to customers. With $2 billion in revenue, the business turns a profit, bringing in more revenue than its costs.

Annual revenue, pre-acquisition

● = $10m

② A private equity fund buys the company, so in addition to its original costs, it has new annual interest payments that cut into its profits.

Annual revenue, post-acquisition

③ As a result of competition from AI, subscription revenue declines. The company's debt burden pushes it into a loss.

Annual revenue, post-acquisition and facing competition from AI

④ To make matters worse, the company's debt is floating rate, so the interest rate it's paying increases. That makes it even harder to cover the firm's original costs and interest payments with existing revenue.

Annual revenue, post-acquisition and facing competition from AI and higher rates

Source: Bloomberg

to Earth' situation that has to happen here in the coming 18 months to two years," says Ruth Yang, head of content strategy at CreditSights, a research service, about the software sector. "If you are a company that is not performing to the standards where you were bought five or six years ago, which is a growing number of companies closely being scrutinized, your ability to find an option other than slamming into the wall at 60 miles per hour is harder."

How Tech Debt Got So Big

A buyout boom was set in motion by the low interest rates that followed the 2008 financial crisis, then the Covid-19 pandemic, which made it much easier for financial magicians to fund acquisitions. At the same time, the post-financial-crisis era of regulation pushed risky lending outside the banking system and encouraged the growth of private credit for huge transactions. Big asset management firms such as Blackstone Inc. and Apollo Global Management Inc., previously known for buyouts, expanded their lending arms. According to the research company Preqin Ltd., assets in institutional private credit funds have grown to about $1.8 trillion globally. Retail investors also began to chase returns by pouring money into private credit funds known as business development companies, or BDCs. Particularly popular was a kind of fund designed to operate perpetually—and always be on the hunt for new capital and deals (second chart on page 68).

Private credit lenders eagerly financed software businesses. While many deals were similar to those in the leveraged loan and junk-bond markets, some acquirers in buyouts began to ask for something more risky: loans based on the promise of contracted subscription revenue, instead of earnings. This was measured by a metric called annual recurring revenue.

One of the investors that made prolific use of these loans was Thoma Bravo, which raised $18 billion in 2020 for a fund that would focus on large software and technology investments. In 2021 the fund purchased Medallia Inc., taking private what was then a publicly traded software company ▶

THE SOFTWARE CREDIT CRUNCH

67

---

focused on online customer surveys. While the company was bleeding cash at the time, it had seen double-digit revenue growth.

Thoma Bravo paid for the riskier equity portion of the $6.4 billion deal, but it turned to a group of private credit lenders, led by Blackstone, to fund the rest with a $1.8 billion loan based on annual recurring revenue, Bloomberg News reported. The lenders were confident enough that they allowed Medallia to delay some interest payments and add to its overall debt, an agreement known as payment in kind. The private credit investors got higher interest rates—and better contractual protections if things went wrong.

In the end, Medallia struggled to generate enough cash, and its debt became untenable. Thoma Bravo handed control of the company to lenders earlier this year—suffering a more than $5 billion hit along with its co-investors, in one of the biggest PE losses since the financial crisis. By the time lenders took over, the loan balance had swelled to about $2.8 billion because Medallia had been delaying some interest payments and borrowed more for add-on acquisitions. BDC funds, which offer a public window into the value of private credit loans, had marked down the value of the debt to as low as 54¢ on the dollar.

Orlando Bravo, the co-founder of the PE firm, told CNBC that his firm “made a mistake” on Medallia by

extrapolating a very high growth rate for the company and paid “too much” for it. He told Bloomberg that other winners in the fund that acquired Medallia would help “bury some losses and some mistakes we made.”

The Risk Around the Corner

Medallia’s troubles predated the recent wave of AI fears, but they point to broader concerns that began to percolate through the lending industry, focused on the outsize levels of debt the software companies were carrying. The interest rates many companies have to pay now are higher than they were expected to be when buyout deals were struck in the early 2020s. More recent deals used less borrowing relative to earnings, or leverage, but they were still premised on rosy revenue growth assumptions. And now markets are signaling a risk of new interest-rate hikes from the Federal Reserve.

Largely as a result of higher interest rates and lower valuations, the private equity industry has been struggling to exit investments—that is, to sell them to other investors or in public stock offerings. This can create a vicious cycle. PE funds under pressure may be less willing or able to kick in more equity to lower the debt burden of their portfolio companies. And then when these businesses need to refinance their debt, they are likely to have to pay an even higher rate, as lenders demand more compensation for

“It is that high financial leverage, the high interest expense, that’s dragging down the overall financial metrics”

Tech Loans Struggle

Bloomberg US Leveraged Loan Index, trading level in cents on the dollar of par, weekly

✓ Main index ✓ Technology sector

Source: Bloomberg

Booming Funds

Business development company assets under management at yearend

■ Listed ■ Unlisted nonperpetual ■ Unlisted perpetual

Source: LSEG BDC Collateral data and company filings compiled by Barclays Research

68

BLOOMBERG MARKETS

---

lending to now-riskier enterprises.

All those problems were there before the AI boom. The big selloff in software leveraged loans began early this year after Anthropic PBC and OpenAI released new models that made it possible for individual users to build programs that could, at least theoretically, replace services provided by many existing software companies. The private credit industry hit its own speed bump as some well-known funds faced redemption requests from their investors.

The moment of truth is expected to come when the software companies need to refinance the debt that was used to finance their buyouts. Across leveraged loans, junk bonds and loans owned by BDCs—a rough proxy for loans made by private lenders—there is more than $150 billion of software company debt coming due between now and the end of 2029, according to research strategists at Barclays Plc. (Unlike with a mortgage, the principal amounts of these loans are due at maturity.) The rule of thumb is to refinance at least one year in advance, preferably earlier, so companies are already bumping against this timeline.

The selloff has hit Perforce Software Inc. hard. Clearlake Capital Group closed its acquisition of the company in 2018, and Francisco Partners joined the investment the next year. Perforce makes software that helps developers build, test and

maintain code—a type of business investors are especially worried will be vulnerable to AI products such as Anthropic’s Claude Code. That concern caused the value of its debt to drop to deeply distressed levels.

There is no indication that customers are using AI to replace Perforce, and the company says that AI should actually help it make even more attractive products. Moody’s Ratings wrote in April that it saw room for revenue growth in 2026, but also that AI adoption “could increase competitive pressure.” And the company has to contend with about $1.7 billion of debt, partly from the initial buyout but also from later acquisitions of other businesses.

Perforce had a $300 million loan coming due in July 2027 and an even bigger loan looming in 2029. Low trading levels would have made a traditional refinancing almost impossible, so instead Perforce struck a deal, Bloomberg reported. Some of its lenders swapped the $300 million debt for notes maturing in 2031, and jumped up in the ranking of who gets paid back first in the event of default.

Another option companies can use is to ask lenders to simply extend the life of the loan. Thoma Bravo-owned Imprivata Inc., which makes identity access management software, did a deal in June that added two years to the debt’s maturity while increasing the interest rate by three-quarters of a percentage point. That was a smaller

Perforce’s AI Pain

Value of Perforce Software’s $1.1b loan due in 2029, in cents on the dollar, weekly

Source: Bloomberg

increase than the company originally asked for, indicating strong demand. Investors saw this as a positive sign for the market.

Still, refinancing can prove tricky even for companies in the cybersecurity sector, which is seen as more insulated from AI. The British company Sophos Ltd., another Thoma Bravo acquisition, tried reaching out to private credit lenders earlier this year to refinance roughly $2.5 billion in leveraged loans that are coming due next year, but several passed, Bloomberg reported. That prompted Thoma Bravo to consider an alternative plan, asking the company’s existing lenders to extend the debt. But they demanded heavy concessions and wanted Thoma Bravo to inject more money into the deal, pay off some existing debt and increase the pricing. In the meantime, the debt is being traded for around 95¢ on the dollar—far from distressed, but a reflection of investors’ uncertainty.

While markets try to figure out just how imminent the threat from AI could be, or if the threat will materialize at all, the refinancing risk remains for indebted private companies. Each update to major AI models reinforces the SaaSpocalypse narrative, and lenders are showing few signs of jumping back in to give the story a happy ending. ●—With Reshmi Basu, Davide Scigliuzzo and Ellen DiMauro

Seligson is a senior reporter on private credit and leveraged finance. Cheng covers leveraged finance.

The Maturity Wall Looms

Software debt coming due, by year

■ US leveraged loans ■ BDC debt investments ■ US high-yield bonds

Maturities for leveraged loans and high-yield bonds are through June 30, 2026; maturities for BDC debt investments are through Dec. 31, 2025. BDCs typically invest in private credit loans but can buy a small portion of other assets such as leveraged loans. Source: Bloomberg and PitchBook data compiled by Barclays Research

THE SOFTWARE CREDIT CRUNCH

69

---

Harol

Europe tries to get savers to take a flier on stocks

---

Cell

Latvia's capital, Riga, at sunset

By LAURA NOONAN and AARON EGLITIS

PHOTOGRAPHS BY REINIS HOFMANIS

---

Toms Kreicbergs worked on Wall Street before returning to his native Latvia, where he writes science fiction and engages in an enterprise almost as fanciful: trying to get more of his fellow citizens to invest in the stock market.

Kreicbergs, who has 211,000 subscribers to his investment advice channel on YouTube, also offers courses whose cost ranges from free to a couple of thousand euros. “In Latvia one person can actually influence public attitudes toward investing,” says Kreicbergs, 40, a onetime commodities banker at UBS Group AG and Scotiabank.

For now, in an age when many around the globe are betting on stocks with a tap on their smartphone, Europeans kept less than 10% of their financial assets in directly held stocks the last time the EU surveyed households. Data shows the proportion in stocks is lower than in the US, where directly held equities counted for about a quarter of financial assets for households and nonprofits.

In countries big and small, the European Union is chasing a dream: Its savers, with €11 trillion ($12.5 trillion) stashed in bank accounts, will finally develop a lust for stock and bond markets and pump money into businesses across the region, powering both European growth and their own prosperity.

Europe is promoting new Savings and Investment Accounts that would let individuals buy shares, bonds and funds for as little as €10 a month. Ideally they’d have some kind of special tax treatment. They could mimic aspects of the tax-free Nippon Individual Savings Accounts, which have driven record retail participation in Japan, as well as Self-Invested Personal Pensions in the UK and 401(k) accounts in the US.

Eleven countries in Europe already have such accounts, and an additional seven, including France, Portugal and Spain, are planning to start them under the initiative. Ideally the funds would have 70% of assets in European stocks, so the investments improve the region’s economic prospects.

“It is a long-term effort to change the way Europe finances its economy and the way Europeans relate to saving and investing,” says Maria Luís Albuquerque, the EU’s financial-services commissioner charged with delivering the plan. “Meaningful progress is already underway, even if change will not happen overnight.”

The EU has also proposed giving enhanced power to the European Securities and Markets Authority. By 2027 or ’28 it would oversee much of Europe’s capital markets, making them more efficient and potentially increasing investor confidence.

But, for every step forward, there’s a step back. The Netherlands unveiled an investment tax in May that will claim

Socking It Away

Share of households’ total financial assets in surveyed European countries

Assets in shares are reported separately from assets in mutual funds and voluntary pensions. Surveyed countries include the 20 euro-area member states as of 2023, plus the Czech Republic and Hungary.

Source: Eurosystem’s Household Finance and Consumption Survey, 2023 wave

72

BLOOMBERG MARKETS

---

Kreicbergs at his recording studio in Riga

36% of paper profits on stocks, bonds and cryptocurrency, even if all the money remains invested. After a backlash, the government said it would soften its impact. Latvia last year increased its capital-gains tax on investments and other assets to 25.5% from 20%. Romania raised taxes on dividends to 16% from 10% in January. (The US taxes long-term capital gains from zero to 20%, with the rate rising with income.)

Jacques de Larosière, the former French central bank governor credited with the idea for a single European supervisor after the financial crisis, says the real barrier to investing has been low EU interest rates, which have driven capital elsewhere. “We say that we want to have a unified market,” says de Larosière, 96, now an adviser to France’s largest bank, BNP Paribas SA. “We say it all the time. I’m fed up with that. Everybody says it, but while the first hand says that, the second hand organizes the exportation of capital outside the region.”

LATVIA COULD BENEFIT more than most from the EU’s initiative. The country has a population of only 1.9 million, and its per capita gross domestic product is 71% of the EU average. On the shores of the Baltic Sea, the nation shares an

eastern border with Russia. Its cities feature a mix of drab midcentury apartment blocks alongside stunning art nouveau architecture. Latvia regained its independence from the Soviet Union in 1990. Ever since, the country has suffered through financial crises.

In the capital of Riga, a city of almost 600,000, Andris Tauriņš recalls how his parents “lost everything” multiple times during his childhood, as they witnessed bank collapses and a currency crisis after the country broke with Russia. “The problem is psychology, people don’t believe,” says Tauriņš, who works as a lawyer. Still, he’s taken an online investing course with Kreicbergs, the YouTuber and novelist. And in his own portfolio he’s put together a kind of rough index of Latvian stocks in which to invest.

Latvia has one of the lowest public stock ownership rates in the EU. Households place just 1.2% of their financial assets in stocks and 87% in bank deposits, the second-highest proportion in Europe. Kaspars Peisenieks, an independent personal finance and investment adviser, says his homeland lacks “positive experiences with the stock market you can pass down through generations.”

The country started its twist on a 401(k)-style plan in 2001, and by 2025 it had collected €8.8 billion, or 22% ►

HARD SELL

73

---

Kulberga at her office

of GDP, according to the International Monetary Fund. Called the Second Pillar, it's meant to supplement employees' traditional government pensions, which provide fixed incomes to retirees, much as Social Security does in the US.

Unlike voluntary 401(k)s in the US, the Latvian version requires most workers to place 5% of wages in the plan. As in the US, they choose from funds ranging from 100% stocks to 100% bonds. Domestic and international companies, including Swedbank AB and Latvia's Indexo, market actively managed funds that invest globally, as well as offerings that track indexes. Upon retirement, in another contrast with 401(k)s, the money is converted into annuities that make regular payments or transferred into the traditional government pension system, resulting in higher monthly incomes.

Latvia itself has little to offer stock enthusiasts. It has been slower than other eastern European countries to list stakes in state-owned companies on its exchange. The nation also offers tax-free investments in government debt and other savings bonds that woo money away from equity markets. Nasdaq's main Riga stock market, which has long underperformed US stocks, is home to just eight companies. Uldis Cērps, head of the Finance Latvia Association, says the country needs a "bold political plan for IPOs."

In a report last year, economists at the European Stability Mechanism, established to address financial crises, recommended that pension funds in Latvia and the other

Baltic nations increase their small allocations to unlisted companies in their home countries. The goal: to encourage growth and eventual entry into the stock market. They also suggested the funds buy more domestic stocks to improve the markets' liquidity.

The IMF has warned that Latvia needs to promote the special investment accounts to grow, and central bank Governor Mārtiņš Kazāks has embraced them. "An economy without a well-functioning capital market is like an athlete competing on a low-calorie diet," he said in a November speech. "He might hang in there for a while but will not reach its peak. An athlete cannot reach the podium by chicken alone."

VISITORS TO THE AIRPORT in Riga will find many companies selling the appeal of markets. Ads for "Banking & Beyond," from British financial technology company Revolut, hawk digital banking. Swedbank is marketing commission-free trading at the local stock exchange.

In a low-rise office building on Riga's outskirts, Karīna Kulberga is on a mission to change attitudes, one investor at a time. A 42-year-old graduate of the Stockholm School of Economics in Riga, she later joined Swedbank before moving to Mintos, a Latvian firm that provides online investing (and whose office has a foosball table and beanbags). Kulberga says Mintos has sold fractional bonds, which let investors buy

74

BLOOMBERG MARKETS

---

part of each security to make them affordable for those with smaller amounts to invest.

Kulberga founded an online community of 10,500 investors and would-be investors on her SheOwns Instagram page, where she seeks to inspire and equip Latvian women to invest. She grew up with a female role model in finance: Her mother is an accountant. In evangelizing to her audience, typically women age 25 to 50, she points to her own example. “I can be an investor as a professional,” she says, “as a full-time mom of two teenage sons, as somebody who is also doing so many other things, as a woman.” Kulberga is already working on the next generation. Her two kids have put 10% of their pocket money into stocks. One of the highlights of her week is a Sunday evening review of their portfolios.

A few miles down the road, Bruno Paļeļunas represents what Kulberga hopes to achieve. The 20-year-old is studying urban planning in the library of Riga Technical University, where he runs an investment club for college students. He invests in Amazon, Tesla, Bank of America and other familiar US names. He also bought shares in a tiny Austrian bank called Bawag Group AG, whose stock has soared this year.

Still, Paļeļunas has found investing a tough sell. After 50 initial responses, most meetings of his club are down to five or six members. His own parents prefer their five rental properties to stocks and bonds, though he’s still hoping to overcome their resistance to equities. He questions whether

Latvians and residents of other small EU countries will venture outside their comfort zone. “You can understand numbers in any language, but I think it’s going to be really hard to invest in small companies in other countries that I’m not really familiar with,” he says.

Sitting in a cafe in Riga’s business district, Boriss Sadrins, who heads Samsung Electronics Co.’s Baltics e-commerce division, says he once invested in his home country, then thought better of it. Now he bets on exchange-traded funds from other markets that have performed better, such as the US. “Money is separate. Heart is one side, money is a second side,” he says. “The States is the best market in the world.”

Then there are those who don’t invest at all. Roberts Idelsons, chairman of Latvia’s Signet Bank AS, says the EU plan hasn’t taken root in the public imagination. “If we ask any of our clients what they think about the savings and investments union in Europe, they would not have a clue what we are talking about,” he says.

Consider Ieva Sauša, a Latvian who buys and sells iPhones independently for a living. She’s keeping her money in the bank, and she shares past generations’ jaundiced view of the stock market: “I believe it’s like a gamble. At the roulette table, someone has to lose, and it’s not the house.” ● —With Charlie Wells

Noonan, based in London, covers global finance; Eglitis, based in Riga, covers economics and government.

Paļeļunas in the kitchen of his family’s home

---

Ideas

Yale's Martha Gimbel on Taxing AI

Interview by BEN STEVERMAN

ILLUSTRATION BY JAYA NICELY

IF ARTIFICIAL INTELLIGENCE is set to transform the economy, shouldn't it also transform the tax code? At a time when the US government takes in far less revenue than it spends, Democrats, Republicans and even some AI executives are debating a range of ideas to share the wealth this technology may create—and to manage how it could disrupt both the economy and the way taxes are collected.

Lately, Martha Gimbel, executive director of the Budget Lab at Yale University, has spent much of her time on the issue, publishing research on AI's economic impact so far and specific ideas to tax it. She co-founded the lab, which analyzes the fiscal impact of federal policy proposals, in 2024 after a career that included work as an economic adviser in President Joe Biden's White House. Our conversation has been edited for length and clarity.

BEN STEVERMAN: Can you just start by defining the problem? What is it about AI that is causing people to rethink taxes?

MARTHA GIMBEL: [Laughs] That is exactly the issue. People do not agree on what taxing AI means, why we need to tax AI, what we're even talking about. I think some of this is just a response to the stress people are having about AI. A thing we can do is tax it. That sounds great, but no one really knows what that means.

It's also intersecting with the sense that people have that the tax code isn't fair, and that people who make a lot of money or have accumulated large amounts of

wealth are not paying their fair share.

BS: What about the fact that workers can end up paying much higher tax rates than investors and business owners? To the extent AI is a labor-saving technology, it could raise real problems for a system that relies so much on workers—rather than investors—paying taxes.

MG: One thing that is really important to think about is that, in the 19th century, countries had to fundamentally change how they were doing taxation. The old system of tariffs, land taxes and things like that didn't work in a situation where all of a sudden you had bankers, factory owners, etc., making a lot of money. That income wasn't being taxed. So it's important to come at this moment from an open mind.

We tax capital less than labor in this country. Historically, there's been a few reasons for that. One is just that capital is much more mobile. It is harder to tax than labor. And so the revenue-maximizing rate for capital has generally been thought to be lower than for labor. Does this new moment change that calculus in some way? I don't think we know yet.

At the same time, leaving aside the mobility of capital, our tax code with respect to taxing capital is riddled with loopholes. And when you have a tax code that is riddled with loopholes and capital is mobile, it's going to flow to where it can benefit from those loopholes.

In my opinion, that is only going to get worse the more powerful AI gets. If you think any of these tools cannot find every

single possible loophole in the tax code, I have a bridge I would like to sell you. It is very probable that the already existing flaws in our capital-gains and corporate taxation system will only become more painful as the technology develops.

BS: One idea you've looked at is taxing AI activity directly, by taxing tokens, which are a way of measuring the amount of data AI models process. What did you conclude?

MG: The tax nerds talk about a token tax. But a token is not a defined concept. It's not like a gigawatt. It's not a pound. Different labs use different tokenizers. In general, it's a bad idea to tax a unit of something where the person you are taxing controls the definition of that unit.

This gets at the question of: What are we trying to do? Are we trying to slow adoption of AI? If you tax something, people are less likely to use it. This is why we have cigarette taxes, for instance. Or do we want to encourage certain uses of AI—for cancer research as opposed to laying off workers.

You would design a tax differently if you were trying to maximize revenue. All of these different use cases require different ways of setting up the tax. Because people are so focused on the slogan of "taxing AI," they're not really thinking through what they're trying to do.

BS: Another way to tax AI might be to tax energy use. What did you decide when you looked into that?

MG: It doesn't have the measurement problem. We all agree what a gigawatt is. It still raises other questions. Both a token tax

76

BLOOMBERG MARKETS

---

and an energy usage tax raise an issue around international competitiveness. Our taxation of AI could impact the ability of American models to compete with Chinese models.

BS: You and the Budget Lab are keeping a very close eye on the labor market, trying to figure out what this technology is doing to the economy. What's your read on where we are right now?

MG: I think the rhetoric around how much AI is currently impacting the labor market has gotten way out over its skis. People feel like they have to rush to do something. But I don't think right now we're seeing any sign of broad economic or labor market disruptions. That isn't to say it is not going to

happen, and this isn't to say we shouldn't be taking policy action at this time or thinking about what are the right things to do.

BS: An idea intriguing people across the political spectrum is the government taking an ownership stake in AI companies, so taxpayers share in the profits. Bernie Sanders has proposed forcing firms to hand over a hefty stake, while OpenAI Chief Executive Officer Sam Altman has talked with Donald Trump about some kind of voluntary arrangement. Do you think that's worth investigating?

MG: In general, government ownership of private enterprise is something we have stayed away from in the US. It can mean that capital is not flowing to the most productive

parts of the economy. And, from a revenue standpoint, it's not clear why we should expect more revenue from an ownership stake than from a reasonable taxation regime. Also, we have no idea which companies are going to be most profitable or see the most benefits coming out of this technological transition. It's still early.

This is a problem with the intersection of economic policy and AI right now. People have decided what the end outcome is going to look like, and they are proposing policies they think fit those very specific outcomes. You want to be thinking about more flexible approaches to policy that can stretch and bend to whatever version of the world you end up in.

So much of this depends on how consumers respond to the technology. It is most interesting to think about media and the arts right now. There's nothing really holding back the rollout of these tools in those industries apart from consumer sentiment.

In healthcare, you have regulatory and liability concerns. No one's going to sue Vogue Japan if Ariana Grande has six fingers on the cover, which is a thing that happened. But consumer sentiment is really toward "human-made" art. It's a thing people still seem to value and still seem willing to pay for. That of course may change.

BS: What are these flexible approaches we could or should be doing anyway, no matter what course AI takes?

MG: Some of the policies aren't particularly sexy or fun. The No.1 thing people need to start thinking about is reducing complexity in the tax code. We need to think about some of the holes in our system that have allowed a lot of capital income to escape taxation. Things like stepped-up basis [which wipes out taxes owed on capital gains when assets are inherited].

None of these are AI-specific. And if you go back to the changes to the tax code people made during the 19th century, none of it was specific to the Industrial Revolution. People didn't tax factory output specifically. What they did was they passed an income tax. It's really important to not lose track of the forest for the trees. ●

Steverman covers money, people and taxes for Bloomberg in New York.

VOLUME 35 / ISSUE 4

77

---

Cheat Sheet

A Compendium of New and Noteworthy Functions

Exploring Alternative Data

ENHANCEMENTS TO TRY RIGHT NOW

ALTD

Alternative Data Company Analysis, which provides intraquarter insights into company performance based on data such as consumer transactions and foot traffic, has been enhanced with new datasets.

Among them is US drug data from Symphony Health, which enables you to track retail prescriptions and wholesale orders for more than 600 drugs sold in the US across 68 companies. The dataset can help you nowcast company key performance indicators, observe sales across new and existing drugs, identify emerging health categories, monitor competitive dynamics and gain earlier insight into company performance.

Consider, for example, Johnson & Johnson's revenue from Tremfya, a biologic drug used to treat psoriasis and Crohn's disease. On the health products company's second-quarter call, Johnson & Johnson reported that Tremfya quarterly revenue rose more than 70% from the previous year. To compare the drug revenue with Symphony prescription data, go to {JNJ US Equity ALTD } and click on the KPI Correlation tab. Use the Reported KPI drop-down to select Tremfya Revenue and the vs Data Source drop-down to

select Symphony Integrated, which represents an integrated view across both retail and institutional channels. The KPI Correlation dashboard section provides data on the statistical and directional accuracy of the estimates versus the reported metric. ALTD provides more than 1,000 KPI estimates in which Symphony pharmaceutical data have been applied to both individual drug segments and overall revenue.

In addition, ALTD now has data from Apptopia, a mobile consumer intelligence provider that offers aggregated mobile app analytics from more than 15 million devices, covering about 2,000 companies in 160 countries. To track app downloads for predictions markets trading company Kalshi Inc., for example, type "Kalshi" in the command line of a Terminal screen and click on the 1859330D US Equity match in autocomplete. The shortcut is {1859330D US Equity ALTD .

To see the evolution of downloads during the past six months, click on the Trend Analysis tab. Set Growth to None: Actual Value and Period to Daily. Click the 6M button. In the Select Alt Data Metric, tick the box for App Downloads and unselect any other data items. Kalshi app downloads jumped as the FIFA World Cup began in June. You can use the Compare Peers section to add data for other companies such as DraftKings, FanDuel and Polymarket Exchange.

BECO MODELS DRIVERS

Bloomberg Economics has updated its oil-price drivers model, which tracks the contributions of economic demand, oil supply and geopolitical risk to changes in the Brent crude price. The model uses high-frequency moves in oil, equity prices and gold to explain moves in Brent. It now uses the Bloomberg World Airlines Index {WAIRLS Index DES} to obtain a clearer signal in equity prices. To see the oil-price model, go to {BECO MODELS DRIVERS }. For more on the methodology, click on Documentation.

BECO MODELS CBSPEAK

Bloomberg Economics has also updated its high-frequency natural language processing model of Federal Reserve policy sentiment. The new model incorporates more recent training data, providing a clearer signal of relative hawkishness or dovishness in communications by members of the Federal Open Market Committee. To see the model, go to {BECO MODELS CBSPEAK }.

Faster, better answers—24/7. Help> for Bloomberg Analytics

The BLOOMBERG PROFESSIONAL service "BPS", BLOOMBERG TERMINAL and Bloomberg data products (the "Services") are owned and distributed by Bloomberg Finance L.P. ("BFLP"), except that Bloomberg L.P. and its subsidiaries distribute the BPS in Argentina, Bermuda, China, India, Japan, and Korea. Bloomberg Tradedock is provided by Bloomberg Tradedock LLC and its affiliates and is available on the BPS. BLOOMBERG, BLOOMBERG PROFESSIONAL, BLOOMBERG MARKETS, BLOOMBERG NEWS, BLOOMBERG ADV MARKET, and BLOOMBERG TRADEDOCK, BLOOMBERG TELEVISION, BLOOMBERG RADIO and BLOOMBERG.COM are trademarks and service marks of BFLP or its subsidiaries. The information depicted within this document and all related information (whether oral or written) is provided for your information and discussion only. It is not a solicitation or offer to buy or sell any financial instrument, nor is it an invitation or inducement to engage in investment activity. Nothing in this document is designed to be, or should be treated as, advice. If you are in any doubt as to your obligations or work advice, you should consult professional advisers. Bloomberg does not guarantee the accuracy or completeness of the content of this document. Bloomberg does not accept liability for any loss arising from use or reliance placed upon this document and its content. Information provided in this document should be treated as proprietary to Bloomberg. This communication is directed only at market professionals who are eligible to be customers of Bloomberg Trading Facility Limited, a limited company incorporated in England and Wales under Company No. 0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000

---

InteractiveBrokers

IBKR Hedge Fund Clients

Outperformed the S&P 500 in 2025

A leading prime broker with zero private equity exposure

IBKR Hedge Funds

S&P 500

Outperformed the S&P 500

28.91%

2025 Average Return¹

17.9%

2025 Return

Support for best execution, margin rates as low as USD 4.13%², industry-low commissions³, and access to 170+ global markets—all on a single, integrated platform.

IBKR Recognized as a Prominent Prime Broker by Preqin⁴

1 Prime Broker for Hedge Funds with AUM under $50 million

Connect with an IBKR specialist to discuss your hedge fund's needs. Call 855-861-6414 or visit

ibkr.com/performance

07-IB26-1812-US

Trading on margin is only for sophisticated investors with high risk tolerance. You may lose more than your initial investment.

Member - NYSE, FINRA, SIPC – Hedge Funds are highly speculative and investors may lose their entire investment. [1] Past performance is not indicative of future results. Returns shown are based on aggregate data for Interactive Brokers accounts meeting minimum thresholds as of January 1, 2025 ($1,000,000 for hedge fund accounts). Such Hedge Fund Accounts accounted for 44.7% of total Hedge Fund Accounts. Returns include all account segments and product types, net of all commissions and charges. Results may vary significantly among clients. Comparisons to the S&P 500 are for informational purposes only. Investing involves risk, including the potential loss of principal. [2] For complete information, see ibkr.com/compare. [3] According to BrokerChooser.com Online Broker Review 2026. For more information, see ibkr.com/info. [4] Source: Preqin Service Providers in Private Markets 2025 report PREQIN® and BLACKROCK® are trademarks of BlackRock, Inc. or its affiliates, used under license.

---

The FFM Quiz

Interpreting Inflation Signals

By ISABELLA YUAN, CLEMENT THIANG and ARUSHI JAIN

INFLATION AND THE PATH OF interest rates are important keys to economic and investing outcomes. Which Group of Seven central bank is adopting a more dovish tone in its communications? How much are tech companies raising prices on computer software and accessories? Are US consumers raising or lowering their inflation expectations? ¶ Test your knowledge with Bloomberg's FFM quiz. Then follow the steps to find the correct answer—and learn more about the Terminal's data and analytical tools.

1 Which of these central banks has shifted tone in a dovish direction compared with its previous core message?	FIND THE ANSWER Run (ASKB ) and click on Workflows. Select the Bloomberg Example workflow titled Central Bank Tracker. In the Configure Variables box, type "G7" and hit . Click the Run button. In the results, scroll down to the Summary Table and check the Core Message and Tone Shift vs. Prior columns.
☐ European Central Bank
☐ Bank of Canada
☐ None of the above
2 Tech companies are raising prices in the wake of massive investments in artificial intelligence. What's the latest year-over-year growth in inflation for computer software and accessories?	FIND THE ANSWER Type "US CPI analyzer" in the command line of a Bloomberg screen and select the ECAN US CPI match. The shortcut is (ECAN US CPI ). Set the Dimension to Percentage Change. Click on Core Goods. Select Education and Communication Commodities, then click on Information Technology Commodities.
☐ Less than 10%
☐ 10% to 15%
☐ More than 15%
3 The Federal Reserve Bank of New York surveys US consumers monthly about the inflation level they expect one year in the future. How have those inflation expectations changed in the past month?	FIND THE ANSWER Type "inflation worksheet" in the command line and click on the WSL INFLATION match. The shortcut is (WSL INFLATION ). Scroll down to Federal Reserve Bank of New York – Survey of Consumer Expectations and read the %1M change for Inflation Expectations Median 1 Year Ahead Expected Inflation Rate.
☐ Increased
☐ Stayed the same
☐ Decreased

80

BLOOMBERG MARKETS

---

MORNINGSTAR | DBRS

Challenging

the status quo since 1976.

Learn more at dbrs.morningstar.com

---

Cartier

制度与权力批判深度研判：Bloomberg Markets (2026-08-16)

▌ 本日全报思想与文化深思雷达 (Intellectual & Cultural Significance Radar)

在系统通盘梳理本期报纸所涉及的所有国际政治、经济、社会与文化报道后，以下 3 个具体事件在思想史、文化政治与制度伦理层面最值得学者进一步深思：

1. 【统计数据的政治化与“真理”的碎片化】 [F4_28, F4_30]

   • 事件简述：在特朗普第二任政府执政期间，美国劳工统计局（BLS）面临质量确保挑战，且2025年秋季政府停摆导致10月份消费者价格指数（CPI）缺失，使得关税成本的峰值无法被官方记录。
   • 思想文化深思切入点：探讨“国家统计”作为一种现代权力治理术（Governmentality）的脆弱性，以及当客观事实被行政权力或制度瘫痪截断时，公共真理如何被碎片化。

2. 【“K型经济”下的技术红利与生存撕裂】 [F2_19, F1_3]

   • 事件简述：AI与数据中心建设为部分人群带来巨大红利，而伊朗战争导致的能源价格上涨使许多家庭陷入困境，同时AI正在颠覆韩国的职业与社交模式。
   • 思想文化深思切入点：分析技术理性如何通过资本逻辑加剧社会分化，探讨AI作为新生产力在创造财富的同时，如何通过重塑生活世界（Life-world）深化阶层之间的生存撕裂。

3. 【衡量标准的偏差与生产力的认知错位】 [F4_31, F4_32]

   • 事件简述：回顾格林斯潘时代CPI的“向上偏差”以及当前AI编程助手在价格不变但能力倍增时，传统通胀衡量标准可能低估生产力增长的现象。
   • 思想文化深思切入点：反思量化指标在捕捉“质量提升”时的失效，探讨当衡量工具滞后于技术演进时，经济决策如何基于一个“扭曲的现实”而运行。

---

▌ 精选核心专题深度思想论证 (In-Depth Dialectical Monograph)

专题一：统计权的解构与治理术的危机：从“国家真理”到“数据真空”

在现代治理术中，统计数据不仅是经济运行的镜像，更是国家行使主权、定义现实的权力工具。然而，本期报道揭示了一种深刻的制度性危机：美国劳工统计局（BLS）在政治环境与制度动荡（如政府停摆）的影响下，出现了关键数据的缺失 [F4_28, F4_30]。2025年秋季CPI数据的空白，意味着关于“特朗普关税成本是否达到顶峰”这一核心经济事实，在官方记录中被永久性地抹除了。

这种缺失并非简单的技术失误，而是一种权力对“可见性”的操纵。当官方数据不再公布，行政权力在事实上获得了一种解释上的豁免权——因为缺乏基准数据，任何关于政策失败的指责都失去了量化支撑。这种现象揭示了统计学（Statistics）在词源上与“国家”（State）的绑定关系：国家通过量化社会生活来实施管理，而当国家选择性地停止量化，其结果是制造出一种“制度性失明”。

进一步分析可见，这种危机导致了认知层面的错位。正如2021年就业岗位数据的累计上调以及2024-2025年工资单数据的下调 [F4_29 🔍]，数据的滞后与扭曲直接影响了美联储的利率决策。这意味着，当统计权的严谨性被政治波动或行政失效侵蚀时，经济治理的逻辑从“基于事实的响应”转向了“基于偏差的猜测”。在这种语境下，真理不再是公共契约的产物，而变成了权力博弈后的残余。

专题二：K型经济的生命政治学：技术红利与生存危机的共时性

本期封面提出的“K型经济”概念 [F2_19 🔍]，不仅是一个经济学图表，更是一种深刻的社会分层隐喻。在当前的全球地缘政治环境下，财富的分配呈现出一种极端的非对称性：一方面是AI驱动的数据中心建设竞赛带来的巨大红利；另一方面则是伊朗战争干扰石油流动、推高能源价格，使普通家庭陷入生存困境 [F2_19 🔍]。

这种结构揭示了技术理性与生命政治之间的冲突。AI在韩国颠覆职业与约会等基础社会关系 [F1_3 🔍]，在宏观层面被视为生产力的飞跃，但在微观层面，这种飞跃是以解构原有的社会契约和生活模式为代价的。对于处于K型曲线顶端的人群，AI是杠杆；而对于处于底端的人群，能源价格的上涨与AI带来的职业替代则是双重挤压。

从异化理论审视，这种分化导致了“生存体验”的彻底脱节。当金融市场在讨论AI带来的增长乐观情绪时 [F6_34 🔍]，底层民众面对的是具体生活成本的攀升。这种共时性（Synchronicity）——即顶端的繁荣与底端的匮乏在同一时间维度内发生——构成了当代社会最深刻的伦理悖论。生产力的提升（如AI编程助手的能力增强 [F4_32 🔍]）并未自动转化为普遍的福祉，反而通过资本的金融化过程，进一步强化了财富的向心力，使“K型”的分叉在技术加速的推动下变得更加不可逾越。

---

▌ 面向学者的开放性思想追问与研究路标 (Open Horizons for Scholarly Inquiry)

1. 【关于统计主权的危机研究】：当国家统计机构（如BLS）因政治干预或制度失效而产生数据真空时，社会如何重建关于“事实”的公共共识？这种“数据缺失”如何成为一种新型的治理手段？
2. 【关于技术红利分配的伦理研究】：在“K型经济”中，AI带来的生产力提升如何被制度化地转化为少数人的红利而非多数人的福祉？如何构建一套能够衡量“技术进步对社会结构撕裂程度”的指标，以修正单一的GDP增长叙事？

理性与情感辩证深度研判：Bloomberg Markets (2026-08-16)

▌ 本日全报生活世界痛感与情感政治深思雷达 (Lifeworld & Affective Significance Radar)

在系统通盘梳理本期报纸所涉及的所有报道后，以下 3 个具体事件在生活世界痛感、情感政治动员与制度理性冲突层面最值得学者进一步深思：

1. 【AI驱动的韩国“K型”社会分层与房产焦虑】 [F1_3, F2_19]

   • 事件简述：人工智能在韩国颠覆了约会、职业及一切生活方式 [F1_3 🔍]，而整体经济呈现出由彼得·阿特沃特（Peter Atwater）推广的“K型”分化趋势 [F2_19 🔍]：AI与数据中心建设为一部分人带来巨大红利，而另一部分人则陷入困境。
   • 情感政治与伦理深思切入点：技术官僚主义将AI视为生产力飞跃的Logos，但其在生活世界的投射却是极端的阶层焦虑。当AI红利与数据中心建设成为财富增长的引擎 [F2_19 🔍]，非红利人群的异化与被抛弃感在“K型”分化的图表下被具象化。

2. 【美国官方统计数据的“失能”与认知危机】 [F4_28, F4_29, F4_30]

   • 事件简述：美国劳工统计局（BLS）的数据出现严重偏差，如2021年就业岗位累计上调190万个 [F4_29 🔍]，且在2025年秋季政府停摆期间未能公布10月份的CPI数据 [F4_30 🔍]。
   • 情感政治与伦理深思切入点：制度理性的基石——“客观数据”——在行政停摆与方法论滞后中出现断裂。当官方数据无法真实反映生活世界的物价与就业痛感时，社会产生了一种深层的“认知不信任感”，导致公共治理的合法性面临挑战。

3. 【AI时代的生产力衡量悖论】 [F4_31, F4_32]

   • 事件简述：回顾格林斯潘时代对计算机定价的关注 [F4_31 🔍]，当前的AI热潮带来了类似挑战：若编程助手价格不变但能力增强一倍，传统的CPI衡量将低估生产力增长 [F4_32 🔍]。
   • 情感政治与伦理深思切入点：这揭示了量化理性在面对指数级技术进步时的失能。当“质量提升”无法被统计指标捕捉时，经济学意义上的“通胀”与个体体感上的“价值”之间出现了严重的认识论脱节。

---

▌ 精选核心专题理性与情感辩证论证 (The Dialectic of Logos and Pathos Monograph)

专题一：统计理性的崩塌与认知真空：美国经济数据“失能”后的治理危机

当代治理体系的合法性高度依赖于一套被认为客观、中立且连续的统计理性（Institutional Logos）。美国劳工统计局（BLS）的CPI和就业数据不仅是美联储制定利率的基准，更是社会公众感知经济温度的官方尺度 [F4_28 🔍]。然而，本期报道揭示了这一理性架构的剧烈撕裂：2025年秋季的政府停摆导致10月份CPI数据的缺失 [F4_30 🔍]，而2021年至2025年间，就业数据的剧烈上调与下调（每年超过100万个岗位）扭曲了劳动力市场的强度感知并推迟了降息 [F4_29 🔍]。

这种制度理性的“失能”产生了一个危险的认知真空。在生活世界中，这种撕裂表现为一种深刻的焦虑：当民众在现实中感受到物价波动，而官方CPI却因为行政原因缺失或因为统计偏差而无法提供真实反馈时，生活世界的真实痛感与制度端的冷色调报告之间出现了严重的脱节。这种脱节不仅是技术性的，更是政治性的——它意味着决定数亿人生活成本的政策基准，正建立在不完整甚至被扭曲的数据之上。

这种辩证摩擦在于：制度端试图通过“确保统计数据质量”的会议来修补漏洞 [F4_28 🔍]，但这种事后的补救无法消除数据缺失期间造成的社会信任赤字。当官方数据成为可以被政治停摆所抹除的符号时，社会心理的韧性将受到冲击，人们意识到决定其生存成本的不再是透明的统计标准，而是充满不确定性的行政状态。

专题二：AI红利的“K型”分化与技术理性的情感异化

在宏观经济学的制度理性架构中，人工智能（AI）被视为生产力飞跃的核心驱动力。本期报道指出，AI以及数据中心的建设竞赛为一部分人带来了巨大的红利 [F2_19 🔍]。然而，这种技术理性的Logos在社会结构中投射出的却是极具破坏性的“K型”分化趋势 [F2_19 🔍]。这种分化不仅体现在财富的绝对值上，更体现在生活方式的根本性颠覆——AI正在韩国颠覆约会、职业及一切生活领域 [F1_3 🔍]。

这里存在着深刻的辩证摩擦：技术官僚将AI视为提升整体效率的工具，但这种“整体”的提升掩盖了微观个体的剧烈阵痛。在“K型”经济的图谱中，红利获得者与被抛弃者之间形成了一道不可逾越的鸿沟。对于被颠覆的职业者而言，AI带来的不是“效率提升”，而是生存空间的被挤压与身份认同的崩塌。

从社会心理韧性诊断来看，这种撕裂正导致一种深层的异化。当AI红利成为唯一的生存安全感来源时，社会竞争被简化为对技术红利的抢夺，而那些无法进入AI生态的人群则在“K型”曲线的下行分支中感受到被时代抛弃的绝望。这种由技术驱动的财富分化，将原本应服务于人类的生产力工具，转化为了一种强化阶层隔离的情感机制。

---

▌ 面向学者的伦理与社会心理开放性追问与研究路标 (Open Horizons for Ethical & Sociological Inquiry)

1. 【关于“统计缺失”与公共信任的社会学研究】：当政府停摆导致关键经济指标（如CPI）缺失时，这种“制度性失明”如何影响公众对政府治理能力的认知？在缺乏官方数据的情况下，社会如何构建临时的“替代真相”？
2. 【关于AI驱动的“K型”分层与情感异化研究】：探讨AI如何通过重塑职业结构，将技术进步转化为一种阶层隔离的情感机制？这种分化如何影响社会成员的归属感与心理韧性？
3. 【关于技术质量提升与通胀衡量之认识论研究】：针对AI编程助手等产品“价格不变但能力倍增”的现象 [F4_32 🔍]，如何构建一套能够捕捉“质量飞跃”的新型经济衡量体系，以避免制度理性对生产力增长的低估？