“Unintended Consequences provides a provocative interpretation of the causes of the global financial crisis and the policies needed to return to rapid growth. Whether you agree or not, this analysis is well worth reading.” - Nouriel Roubini, New York University; Chairman, Roubini Global Economics
“…a comprehensive explanation of the modern economy.” - Julian Robertson, Founder, Tiger Management
“…challenges misconceptions that distort our economic debates.” - Arthur Brooks, President of the American Enterprise Institute
“Unintended Consequences offers deep and well-argued analyses on almost every issue.” - The New York Times
“…serious thinking for serious thinkers. …a thought-provoking blueprint for growing middle- and working-class incomes.” - Mitt Romney, former Governor of Massachusetts
“…serious thinking for serious thinkers. …a thought-provoking blueprint for growing middle- and working-class incomes.” - Mitt Romney, former Governor of Massachusetts
“A full-throated defense of economic dynamism.” - The Wall Street Journal
“…a fresh argument for the productive value of inequality.” - David Autor, Professor of Economics, Massachusetts Institute of Technology
“…a fresh argument for the productive value of inequality.” - David Autor, Professor of Economics, Massachusetts Institute of Technology
“…a must-read for serious students of economic policy.” - Glenn Hubbard, Dean, Columbia Business School, and former Chairman of the Council of Economic Advisers
“…a must-read for serious students of economic policy.” - Glenn Hubbard, Dean, Columbia Business School, and former Chairman of the Council of Economic Advisers
“…reminds us that inequality sends a signal of what society lacks most, in America’s case, entrepreneurship and risk taking.” - Lawrence Lindsey, CEO, The Lindsey Group, former Director of the National Economic Council
AI Summary.A study of US private firms finds that a one-standard-deviation rise in a firm's financing shortfall is associated with a 45% higher likelihood of going public. After an initial public offering, disagreement among lenders about a firm's default risk falls, borrowing costs drop by 45 basis points, and debt rises more than 50% relative to comparable firms by year three.
Andres Almazan, Nathan Swem, Sheridan Titman and Gregory WeitznerNational Bureau of Economic Research
Date Posted:
A study of 100k+ private nonfinancial firms (2012–23), including 433 IPOs, suggests listing improves access to debt as well as equity. After listing, lenders’ risk assessments converge and bank borrowing costs fall 45 basis points, controlling for assessed risk.
Do financing shortfalls drive private firms to go public?
Core argument: A one-standard-deviation increase in a US private firm’s financing deficit is associated with a 45% higher likelihood of going public, controlling for size, growth, leverage, investment, and industry–location–time effects.
A one-standard-deviation increase in a firm’s financing deficit is associated with a 45% higher likelihood of going public [controlling for size, growth, leverage, investment, and industry–location–time effects]. Consistent with a reduction in asymmetric information, we find that dispersion in PD [probability of default as perceived by disparate market participations] drops,beginning immediately after the IPO. [Note that] we calculate PD dispersion only for banks present prior to the IPO]. We find that after controlling for these risk assessments, firms’ borrowing costs drop by 45 bps after going public. [controlling for bank–firm and bank–quarter fixed effects]. Figure 7 shows that in the first year, there is a statistically significant drop in leverage of about 4 pp.. However, the point estimate returns to zero by year three. Figure 6 shows that IPO firms increase their debt by more than 50% relative to control firms by year three following the IPO. [An important caveat is that our post-IPO estimates capture both treatment and selection effects].
Takeaways by Macro Roundup® AI
A one-standard-deviation increase in a US private firm’s financing deficit is associated with a 45% higher likelihood of going public, controlling for size, growth, leverage, investment, and industry–location–time effects.
Dispersion in the probability of default perceived by banks present before the IPO drops immediately after it, and after controlling for these risk assessments, firms’ borrowing costs fall 45 basis points following the IPO, with bank–firm and bank–quarter fixed effects.
IPO firms’ leverage falls a statistically significant 4 percentage points in the first year, with the point estimate returning to zero by year three, while their debt rises more than 50% relative to control firms by year three; these post-IPO estimates capture both treatment and selection effects.
Related Articles:
A Catalyst For The AI Bubble Break— Historically, controlling for valuations, a 1% rise in the share of the market represented by initial public offerings has been associated with roughly a 4% decline in prices. With large artificial-intelligence company listings expected and share issuance exceeding buybacks, equity supply could shift from shrinking about 1% a year to growing near 5%, implying returns roughly 20% below average.
The Coming Mega-IPO Flow & Funding Problem of 2026— Adjusted for inflation, SpaceX alone would rank as the second-largest initial public offering ever, behind Saudi Aramco, and three such listings together would exceed the entire 1995–2000 dot-com wave. Funding that equity supply requires selling existing holdings, and index rule changes make passive funds forced buyers, pressuring large-cap technology stocks.
The Deep End: 2025 Alternative Investments Review— In venture capital, only consistently investing with top-quartile managers outperformed the S&P 500, while the median manager at best matched public equity markets and fourth-quartile managers destroyed value relative to equities. Fees are part of the issue: an 11.7% annualized NASDAQ 100 return falls to 7.4% under a 2%/20% fee structure.
AI Summary.China's exports rose 36% from a year earlier and imports 31%, widening its monthly trade surplus to almost $134bn and putting the annual surplus on track to exceed the $1.2tn record. High-tech shipments jumped almost 64%, which Bloomberg Economics estimates accounted for about half of total export growth.
James Mayger and Paul AbelskyBloomberg
Date Posted:
China’s goods trade surplus is on track to exceed 2025’s $1.2T. China’s trade surplus with the EU is up 21% y/y.
Is China's trade surplus heading to a new record on export growth?
Core argument: China’s exports rose 36% year-over-year in September, the fastest pace since February, while imports rose 31%, widening the monthly trade surplus to almost $134 billion and exceeding every forecast in a Bloomberg survey of economists.
China’s export growth surged the most since February to drive global imbalances to new highs, with its trade surplus this year on track to exceed the $1.2 trillion record set in 2025. Shipments abroad soared 36% last month from a year earlier and imports jumped 31%, data released by China’s General Administration of Customs showed on Wednesday. As a result, the surplus for September widened to almost $134 billion, exceeding every forecast in a Bloomberg survey of economists. The imbalance has continued to balloon since US President Donald Trump ramped up tariffs soon after reclaiming the White House at the start of last year, showing the limits of the policy in the middle of the AI buildout. Shipments of high-tech products soared almost 64% in September, which Bloomberg Economics estimates contributed around half of China’s total export growth.
Takeaways by Macro Roundup® AI
China’s exports rose 36% year-over-year in September, the fastest pace since February, while imports rose 31%, widening the monthly trade surplus to almost $134 billion and exceeding every forecast in a Bloomberg survey of economists.
China’s trade surplus this year is on track to exceed the $1.2 trillion record set in 2025, with the imbalance continuing to grow after US President Donald Trump raised tariffs at the start of last year.
Chinese shipments of high-tech products soared almost 64% in September, contributing around half of total export growth, according to Bloomberg Economics estimates.
Related Articles:
The World Should Not Ignore China’s Undervalued Currency— China's reported current account surplus of about 4% of GDP is roughly 3 percentage points above the International Monetary Fund's estimated norm, implying a currency undervaluation of about 20%. Adjusting for surging gold imports and a misreported investment income deficit would lift the surplus to about 6% of GDP and the estimated undervaluation to 30–35%.
World Economic Outlook: Resilience Under Strain, Urgent Choices— Global current account balances widened to about 3.8% of world GDP, driven mainly by a rise in China's surplus to about 0.6% of world GDP (3.7% of China's GDP) and larger surpluses across other Asian economies, while the United States remained the largest deficit economy at about 1% of world GDP. Imbalances are projected to remain elevated.
Understanding Global Imbalances— The US, China, Germany, and Japan account for roughly two-thirds of global current account imbalances, with the US deficit at 4% of GDP financed by capital inflows into dollar assets. The average duration of a surplus or deficit spell roughly doubled since the 1980s, building large net foreign asset and liability positions.
AI Summary.US consumer prices rose 3.7% year over year in September, up from 3.4% in August and above economists' 3.6% forecast, with core prices excluding food and energy up 2.6%. Prices rose faster than average hourly earnings, which increased 3%, meaning many Americans' real incomes are falling.
Konrad PutzierWall Street Journal
Date Posted:
Inflation accelerated in September, rising to 3.7% annualized from August’s 3.4%. Real average hourly earnings fell .7% year over year.
Is inflation accelerating faster than wages, cutting real incomes?
Core argument: U.S. consumer prices rose 3.7% in September from a year earlier, the Labor Department said, up from August’s 3.4% and above the 3.6% expected by economists surveyed by The Wall Street Journal.
Consumer prices were up 3.7% in September from a year earlier, the Labor Department said Wednesday, hotter than August’s gain of 3.4%. That was above the 3.6% rise expected by economists surveyed by The Wall Street Journal. Core prices excluding food and energy rose 2.6%, above forecasts for a 2.5% increase. Month over month, core prices rose by 0.3%, surprising economists who had expected a 0.2% increase. High prices have been a political flashpoint for Americans, who have grown fed up with five years of inflation. Prices rose much faster in September than average hourly earnings, which increased 3% from a year earlier. That means many Americans are seeing their real incomes fall.
Takeaways by Macro Roundup® AI
U.S. consumer prices rose 3.7% in September from a year earlier, the Labor Department said, up from August’s 3.4% and above the 3.6% expected by economists surveyed by The Wall Street Journal.
Core prices excluding food and energy rose 2.6% year over year in September, above forecasts of 2.5%, and 0.3% month over month against expectations of 0.2%.
September prices rose much faster than average hourly earnings, which increased 3% from a year earlier, meaning many Americans are seeing real incomes fall.
Related Articles:
Do Voters Punish Inflation or Pay Cuts? Inflation and Real Wages in U.S. Elections— An analysis of U.S. counties estimates that where wages grew faster than prices, voters supported the incumbent party more, while counties with larger real wage losses than average shifted more toward the opposition. A one-standard-deviation larger real wage loss is associated with shifts of roughly 0.10 and 0.19 percentage points in vote share and margin.
Sticky Wage Norms and the Real Wage Cost of Unexpected Inflation— Firms set nominal wage increases by round-number rules—about 42% of increases below 6% fell within 0.01 percentage points of a whole or half number—and these rules rose only from about 2.7% to 3.5% as inflation peaked near 7%, contributing to falling real wages for job stayers.
The Cost of Money is Part of the Cost of Living: New Evidence on the Consumer Sentiment Anomaly— Variation in the University of Michigan consumer sentiment index that inflation and unemployment cannot explain is strongly correlated with growth in consumer borrowing costs, such as mortgage and car loan rates. The authors estimate that including interest payments and homeownership costs in the consumer price index would raise measured inflation from 3% to 9%.
AI Summary.A survey of 396 enterprises finds only 11% forecast their AI spending within 10% of actual, down from 15% a year earlier, while 62% say an unexpected AI cost altered a business decision and one in four delayed or canceled an AI initiative. 81% report moderate-to-high gross margin impact from AI costs.
Torsten SlokApollo
Date Posted:
Slok flags a survey of 396 enterprises that found AI adaptation hasn’t improved gross margins in aggregate, with 81% of firms noting a “moderate-to-high gross margin impact” from their AI spending, while 63% saw a 6 to 15pp decrease.
Are companies losing control of their AI spending forecasts?
Core argument: Only 11% of the 396 enterprises surveyed by Mavvrik and Benchmarkit forecast their AI spending within 10% of actual, down from 15% a year earlier.
A new survey by Mavvrik and Benchmarkit of 396 enterprises finds that companies are adopting AI faster than they are building the financial controls to keep its costs in check. Only 11% of firms forecast their AI spending within 10% of actual, down from 15% last year. The survey also shows that those misses are changing decisions, with 62% of firms saying an unexpected AI cost altered a business decision and one in four delaying or canceling an AI initiative. The impact is showing up in profitability too, with 81% of companies reporting moderate-to-high gross margin impact from AI costs, including 63% who see a 6 to 15 percentage point decrease.
Takeaways by Macro Roundup® AI
Only 11% of the 396 enterprises surveyed by Mavvrik and Benchmarkit forecast their AI spending within 10% of actual, down from 15% a year earlier.
In the Mavvrik and Benchmarkit survey, 62% of firms said an unexpected AI cost altered a business decision, and one in four delayed or canceled an AI initiative.
The survey reports that 81% of companies see moderate-to-high gross margin impact from AI costs, including 63% citing a 6 to 15 percentage point decrease.
Related Articles:
Tech’s Trillion-Dollar Internal Inconsistency— Wall Street analysts covering technology project the sector's operating cash flow will more than double to roughly $2.4tn by 2028, a gain of over $1.2tn, while analysts covering the rest of the S&P 500 — technology's customers — project those firms will add far less cash flow. These sector-by-sector forecasts are internally inconsistent.
How Much Revenue Does The AI Buildout Need In 2030?— The authors estimate AI capital spending of $560bn in 2026 and, in a high-investment case, $3.5tn of net AI capital in service by 2030, requiring roughly $1.3tn in annual infrastructure revenue once depreciation, operating costs and a 15% return are covered. That implies 7.4x growth from current annualized revenues, with downstream revenues of $1.9–2.6tn.
Writing Code vs. Shipping Code: Productivity Effects Across Generations of AI Coding Tools— The authors estimate that AI coding tools raise developers' task-level productivity by about 40% with autocomplete, 140% including synchronous agents, and 180% with asynchronous agents, with larger effects for less active developers. Gains attenuate up the production chain: synchronous agents raise lines of code 741% and pull requests 65% but releases only 20%.
AI Summary.Across 12 middle-income countries, call-center employment grew in line with other white-collar office work before the spread of AI chatbots but has since stagnated or declined even as other white-collar jobs held up. The authors argue this divergence makes artificial intelligence's role hard to dismiss.
Caelan Wilkie-RogersRevelio Labs
Date Posted:
A Revelio Labs analysis finds global call center employment is now 5.3% under its December 2023 peak. Call center employment started diverging from other white collar work in middle income countries around November 2022, when ChatGPT was released.
Is AI reducing call center employment in middle-income countries?
Core argument: Across 12 middle-income countries, call-center employment showed sluggish growth or outright declines after ChatGPT’s release, while other white-collar office and administrative roles remained fairly robust.
Across the 12 middle-income countries, the pattern is recognizable: in the pre-ChatGPT era, call-center headcount generally kept pace with other office and administrative, "white-collar" roles. In the post-ChatGPT era, we see sluggish growth or outright declines in call centers—even as the rest of white-collar work looks fairly robust. So, although we cannot observe which firms replace people with chatbots, this divergence in headcounts between call centers and the rest of their labor markets makes AI’s potential role in this trend hard to wave off. Moreover, call centers’ high-volume, scripted work is exactly the profile the International Labour Organization flags as most exposed in these countries. The industry is shrinking because hiring slowed, not because separations picked up.
Takeaways by Macro Roundup® AI
Across 12 middle-income countries, call-center employment showed sluggish growth or outright declines after ChatGPT’s release, while other white-collar office and administrative roles remained fairly robust.
Before ChatGPT, call-center headcount in these 12 middle-income countries generally kept pace with other office and administrative employment.
The ILO flags high-volume, scripted work—the profile of call-center jobs—as most exposed in these middle-income countries, though which firms replace workers with chatbots cannot be observed.
Related Articles:
AI’s Most Exposed Industry Keeps Hiring— Call center employment in the Philippines has reached nearly 2m workers and has grown every year since 2016, including through the artificial intelligence boom. As AI lowers the cost per customer interaction, companies are buying more of these services rather than fewer — an example of Jevons paradox.
Looking for the Ladder— The article argues that the hiring downturn in occupations most exposed to generative AI reflects interest-rate sensitivity rather than AI itself, since roughly 38% of workers in the most AI-exposed quintile work in information, finance and insurance, and professional and technical services, versus under 2% in the least-exposed quintile.
The Early Impacts of AI on Employment among Recent College Graduates— An expanded unemployment measure for young college graduates — including those outside the labor force who report wanting a job — averaged 10.4% in the most recent summer, the highest of the five summers compared and 0.3 percentage points above the prior high. Estimated associations with occupational AI exposure are positive but not statistically significant.
AI Summary.The United States generates almost twice as much economic output per dollar invested as China, the opposite of what would be expected from a mature developed economy versus a fast-growing economy catching up. The gap points to overexuberance in Chinese investment.
Robin Rivaton and Ben SouthwoodWorks In Progress
Date Posted:
The US generates ~2x as much output per dollar of capital as China. Rivaton and Southwood argue that relentless competition among Chinese subnational governments, who account for ~85% of public expenditure, drives industrial growth at the cost of reduced capital efficiency.
Does China get less economic output per dollar invested than America?
Core argument: The United States obtains nearly twice as much GDP from every dollar invested as China does, measured by the capital efficiency ratio of output to invested capital stock.
Overexuberance is visible in economic data. Interest rates that banks pay to depositors are capped, giving banks artificially cheap capital. Entrepreneurs may be backed to attempt extremely unlikely, borderline schemes. The capital efficiency ratio measures how much output a country gets out of the capital stock it has invested – how many dollars of GDP a country can buy with a dollar of investment. The United States gets almost twice as much economic output out of every dollar it invests as China does, precisely the opposite of what we would expect from a developed country with a mature economy versus a rapidly growing country playing catch up.
Takeaways by Macro Roundup® AI
The United States obtains nearly twice as much GDP from every dollar invested as China does, measured by the capital efficiency ratio of output to invested capital stock.
China’s capital efficiency inverts the expected pattern, in which a rapidly growing catch-up economy would extract more output per dollar of investment than a mature developed economy.
Overexuberance is visible in Chinese macroeconomic data, with China’s investment generating less economic output per dollar than that of the United States.
Related Articles:
The Year of the Trojan Fire Horse: China’s Imbalanced Economy and Unrelenting Mercantilism— About 30% of Chinese industrial companies are unprofitable "zombies" kept alive by subsidies far larger than those elsewhere, rather than exiting amid falling industrial producer prices. Countries with large increases in manufactured imports from China also saw slowdowns in domestic manufacturing, including declining manufacturing shares of output across most of Southeast Asia.
Xi’s Superpower Ambitions Fuel a Dangerous Debt Spiral in China— China's credit to the nonfinancial sector—corporate, government and household borrowing—topped 300% of gross domestic product, 60 percentage points above its end-2019 level and larger than in the US, according to the Bank for International Settlements. The International Monetary Fund estimates broad Chinese government debt, including local financing vehicles, reached 127% of GDP, up from 98% in 2020.
Is China’s High-Quality Investment Output Economically Viable?— China's technological progress and infrastructure do not refute claims of overinvestment, since overinvestment often produces real physical assets and technological gains yet still counts as overinvestment when it generates less real value than the resources, subsidies, and transfers it consumes. When such losses are capitalized, assets are carried above their real economic value.
AI Summary.Global current account balances widened to about 3.8% of world GDP, driven mainly by a rise in China's surplus to about 0.6% of world GDP (3.7% of China's GDP) and larger surpluses across other Asian economies, while the United States remained the largest deficit economy at about 1% of world GDP. Imbalances are projected to remain elevated.
IMF StaffInternational Monetary Fund
Date Posted:
Global current account balances increased further in 2025 to ~3.8% of global GDP. The IMF calculates the rise was driven by a sharp increase in China’s surplus to ~.6% of global GDP (3.7% of China’s). The US deficit was ~1% of global GDP.
Are global current account imbalances widening as China's surplus grows?
Core argument: Global current account balances widened to about 3.8% of world GDP in 2025, continuing the upward trend since the pandemic and reversing the post-global-financial-crisis decline.
Global current account balances increased further in 2025 to about 3.8% of world GDP, continuing the upward trend observed since the pandemic and reversing the decline following the global financial crisis. The widening was driven primarily by a sharp increase in China’s surplus, which increased to about 0.6% of world GDP (3.7% of China’s GDP), alongside larger surpluses in several Asian economies, whereas the euro area surplus narrowed and oil exporters’ surpluses declined. At the same time, the United States remained the world’s largest deficit economy, with a current account deficit of about 1% of world GDP. Early 2026 trade data indicate that global imbalances are expected to increase slightly this year and are projected to remain elevated in 2027, with further increases in trade surpluses in China and other Asian exporters and weaker balances in the United States and the euro area.
Takeaways by Macro Roundup® AI
Global current account balances widened to about 3.8% of world GDP in 2025, continuing the upward trend since the pandemic and reversing the post-global-financial-crisis decline.
China’s surplus rose to about 0.6% of world GDP (3.7% of China’s GDP), driving the widening alongside larger surpluses in several Asian economies, while the euro area surplus narrowed and oil exporters’ surpluses declined.
The United States remained the world’s largest deficit economy at a current account deficit of about 1% of world GDP, and early 2026 trade data indicate global imbalances are expected to increase slightly in 2026 and remain elevated in 2027.
Related Articles:
Understanding Global Imbalances— The US, China, Germany, and Japan account for roughly two-thirds of global current account imbalances, with the US deficit at 4% of GDP financed by capital inflows into dollar assets. The average duration of a surplus or deficit spell roughly doubled since the 1980s, building large net foreign asset and liability positions.
Who Will Win The War Of Neo-Mercantilists?— China's manufacturing export surplus is around 2% of world GDP, roughly twice the largest surplus Japan ever ran, while domestic demand has stalled. Unexplained negative net income of about $125bn on its $4tn in net foreign assets, versus an expected $100bn surplus, supports the argument that the renminbi is about 30% undervalued.
The U.S. Trade Deficit: Myths and Realities— Taxing capital inflows would weaken the dollar but could raise domestic interest rates and depress investment, harming long-term growth, while an interest rate cut absent recession would be inflationary and erode any real depreciation needed to durably improve the trade balance. Fiscal restraint could weaken the dollar and reduce the trade deficit.
AI Summary.Goldman Sachs estimates that artificial intelligence–related stocks account for 57% of Chinese equity trading turnover and that nearly $5tn in AI-related market value has been created, including $2tn from new listings. China also holds dominant shares of the physical supply chain, including 97% of humanoid robot shipments, 75% of electric vehicle production, and 90% of rare-earth refining.
Paul KedroskyApplied Complexity
Date Posted:
China accounts for 77% of global token usage, but only 22% of token spending, according to a GS analysis. Kedrosky notes, “This is what widespread AI adoption at low prices looks like. And it is the channel through which Chinese competition will pressure prices elsewhere.”
How large is AI's footprint in Chinese markets and supply chains?
Core argument: AI-related stocks account for 57% of trading turnover in Chinese equities, and Goldman estimates almost $5 trillion in AI-related market value has been created since the DeepSeek moment, including $2 trillion from IPOs.
AI-related stocks now represent 57% of trading turnover in Chinese equities. Since the DeepSeek moment, Goldman estimates almost $5 trillion in AI-related market value has been created, including $2 trillion from IPOs. AI dominates Chinese equity markets even more than it does U.S. markets. It's a one-factor world. China shipped 97% of humanoid robots in the first half of 2026. Alongside China’s 75% share of EV production and 80% of battery production, China basically owns the physical side of AI and AI-adjacent technologies. The industrial base behind the AI buildout is already concentrated in China: 60% of global transformer production capacity, 60% of PCB and optical-module production, 77% of new foundry capacity, and 90% of rare-earth refining. There are more than 880 AI-related companies focused mainly on China’s domestic market, with $3.2 trillion in market value. To put it in context, this is nearly half the total market of the Nasdaq at the peak of the dot-com frenzy.
Takeaways by Macro Roundup® AI
AI-related stocks account for 57% of trading turnover in Chinese equities, and Goldman estimates almost $5 trillion in AI-related market value has been created since the DeepSeek moment, including $2 trillion from IPOs.
China shipped 97% of humanoid robots in the first half of 2026, alongside a 75% share of EV production and 80% of battery production.
China holds 60% of global transformer production capacity, 60% of PCB and optical-module production, 77% of new foundry capacity, and 90% of rare-earth refining, concentrating the industrial base behind the AI buildout.
Related Articles:
Who’s Afraid of Chinese Models?— Demand for frontier artificial intelligence models exceeds supply because computing capacity is scarce, allowing chip suppliers and their customers to resell compute at high margins. The author argues alarm over Chinese models is overblown, expecting the inference market to grow faster than training costs so labs can thrive at lower prices.
Semiquincententacles— Leading artificial-intelligence labs are moving to usage-based token pricing on a cost-plus basis that includes total ownership costs, pushing token prices higher and giving some enterprises reason to migrate to cheaper open models. Frontier models remain necessary for cybersecurity, distillation training, agents, and scientific discovery, where maximum performance is required.
Why .400 Hitters Disappeared — and What It Means for AI— Artificial intelligence models' capability gains are shrinking in percentage terms, falling from about 15% relative to the baseline early on to roughly 10% for the same absolute increment, narrowing the gap between frontier and good-enough models. The author argues this points to a commoditizing market where price, not minor model differences, becomes the main differentiator, pressuring frontier firms' margins.
