Edward Conard

Top Ten New York Times Bestselling Author

  • “…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 fresh argument for the productive value of inequality.” - David Autor, Professor of Economics, Massachusetts Institute of Technology
  • “…a comprehensive explanation of the modern economy.” - Julian Robertson, Founder, Tiger Management
  • “Unintended Consequences offers deep and well-argued analyses on almost every issue.” - The New York Times
  • “…challenges misconceptions that distort our economic debates.” - Arthur Brooks, President of the American Enterprise Institute
  • “Unintended Consequences is far smarter and more thought-provoking than most economics written for the general public” - Greg Mankiw, Harvard University, Former Chairman of the Council of Economic Advisors
  • “…serious thinking for serious thinkers. …a thought-provoking blueprint for growing middle- and working-class incomes.” - Mitt Romney, former Governor of Massachusetts
  • “…challenges misconceptions that distort our economic debates.” - Arthur Brooks, President of the American Enterprise Institute
  • “…a fresh argument for the productive value of inequality.” - David Autor, Professor of Economics, Massachusetts Institute of Technology
  • “A full-throated defense of economic dynamism.” - The Wall Street Journal
  • “…serious thinking for serious thinkers. …a thought-provoking blueprint for growing middle- and working-class incomes.” - Mitt Romney, former Governor of Massachusetts
  • “…a very valuable contribution.” - Larry Summers, former Secretary of the Treasury and director of the National Economic Council, president emeritus, Harvard University
Upside of Inequality Oxford Unintended Consequences

Macro Roundup


Friday, August 8, 2025

AI Summary. Data center capital spending differs from fiber and railroad investment because most of the cost sits in processors that are replaced every 3-4 years rather than in long-lived shells, power, and cooling connections. That short replacement cycle requires returns high enough to exceed a risk-adjusted cost of capital of 12-14% before the hardware must be replaced.

Paul Kedrosky Applied Complexity
Date Posted:
Is Database:
Database

Paul Kedrosky estimates given datacenter’s high deprecation rates – the replacement cycle for GPUs is 3 to 4 years – to be profitable, the return “must exceed the risk-adjusted, weighted average cost of capital, which is 12–14%.”

Does data center spending resemble investment in fiber and railroads?

Core argument: GPUs face a replacement cycle of 3–4 years at best, unlike long-lived assets such as backbone fiber and railroads, making data center capital expenditure different from those earlier buildouts.

Capital expenditure on data centers is very different from spending on fiber and railroads. Railroads and backbone fiber are long-lived assets. This is not the case with data centers. Almost all aspects have declining utility over time. First, and most obviously, the GPUs running and being used for training or inference purposes. The replacement cycle for GPUs is 3-4 years, at best, nothing like fiber or railroads. To turn it around, if you create a data center today and don't use it for four years, it would be better to never turn on the power. It's largely been superseded by subsequent technologies. Data centers also require power and water, and those connections, within reason, are useful longer. The outer shell—some would call it a building—also has a longer life. But most of the cost in a data center is not in the shell, or power, or cooling water. It is in the processors. And having to replace them every few years creates intense pressure on the investment. You must earn a high enough return before replacement to justify the expenditure. In financial terms, your income must exceed the risk-adjusted, weighted average cost of capital, which runs 12-14%.

Takeaways by Macro Roundup® AI

  1. GPUs face a replacement cycle of 3–4 years at best, unlike long-lived assets such as backbone fiber and railroads, making data center capital expenditure different from those earlier buildouts.
  2. Most of a data center's cost lies in the processors rather than the outer shell, power connections, or cooling water, which have longer useful lives.
  3. Data center investment must earn income exceeding a risk-adjusted weighted average cost of capital of 12–14% before processors are replaced, and a facility left unused for four years would be better never powered on.

Related Articles:

  • Back to Our Regularly Scheduled Programming — The four largest US hyperscalers spent $440bn on capital expenditure and research in 2024, with $596bn projected for 2025, a 35% increase, leaving large gaps between AI spending and AI revenues. These firms account for more than half of all capital expenditure growth in the US equity market.
  • Honey, AI Capex Keeps Eating … Everything — The author estimates that AI-related capital spending contributed between 0.6% and 1.3% of a quarter's 3% U.S. GDP growth, based on information technology equipment and intellectual property investment data. Google, Amazon, Meta, and Microsoft alone spent about $69bn in the quarter, roughly half of all information technology equipment spending on an annualized basis.
  • Honey, AI Capex is Eating the Economy — Capital spending on AI data centers is around 20% of railroads' peak spending as a share of GDP and is rising quickly, having already surpassed the dot-com era peak in telecom investment. Unlike long-lived rail infrastructure, data centers are short-lived assets requiring frequent hardware replacement, and the capital is being reallocated from other sectors.

AI Summary. Nvidia holds the largest weight of any individual stock in the S&P 500 since the data series began in 1981, reflecting extreme concentration in AI-linked shares.

Torsten Sløk Apollo
Date Posted:

Nvidia’s 8% weight in the S&P 500 is the largest of any individual stock since the start of the data in 1981. Nvidia has a P/E ratio of 58.

Has stock market concentration reached its highest level on record?

Core argument: Nvidia holds the largest weight of any individual stock in the S&P 500 since the data series began in 1981, according to a chart of the index's biggest company by market capitalization.

The chart shows the biggest stock by market cap in the S&P 500, and it confirms the extreme AI concentration in the market today. Nvidia now has the biggest weight in the S&P 500 of any individual stock since the data began in 1981.

Takeaways by Macro Roundup® AI

  1. Nvidia holds the largest weight of any individual stock in the S&P 500 since the data series began in 1981, according to a chart of the index's biggest company by market capitalization.
  2. The S&P 500's largest-stock weighting points to extreme AI concentration in the market, with Nvidia exceeding every prior single-stock level recorded since 1981.

Related Articles:

  • Summer Mailbag — The share of total US market capitalization held by young unprofitable companies has fallen from its 2021–2022 peak, while over 40% of small-cap companies have negative earnings and the negative-earnings share among large caps is falling. The 10 largest stocks hold about 35% of market cap, a high since 1990, but their share of market earnings has roughly kept pace.
  • Which U.S. Stocks Generated the Highest Long-Term Returns? — Of the 29,078 U.S. common stocks in a database spanning nearly a century, 51.6% had negative cumulative returns, while 17 stocks returned more than 5m% — led by Altria at 265m%. Those top performers averaged annualized returns of just 13.47%, underscoring the value of time invested.
  • Long-Term Shareholder Returns: Evidence From 64,000 Global Stocks — A study of about 64,000 global stocks estimates net global stock market wealth creation of $75.7tn, with returns highly concentrated: the top 159 firms (0.25%) account for half of net wealth creation and the top 1,526 (2.39%) account for all of it. Fewer than half of stocks beat one-month US Treasury bills, 42.6% of non-US and 44.8% of US stocks.

AI Summary. Survey data show adults in their twenties and thirties increasingly report being easily distracted, careless, less tenacious, and less likely to make and keep commitments, a decline in conscientiousness. The same data show rising neuroticism and lower sociability, with young adults — long the most extroverted group — now reporting the most introversion.

John Burn-Murdoch Financial Times
Date Posted:
Is Database:
Database

Btw 2014 and 2024, there was a “sustained erosion of conscientiousness” in young Americans (ages 16–39); they report “feeling increasingly easily distracted and careless, less tenacious and less likely to make and deliver on commitments.”

Are young adults becoming less conscientious, more anxious, and more introverted?

Core argument: People in their twenties and thirties report being increasingly easily distracted and careless, less tenacious, and less likely to make and deliver on commitments, according to the Understanding America Survey.

Digging deeper into the data, which comes from the Understanding America Survey, we can see that people in their twenties and thirties in particular report feeling increasingly easily distracted and careless, less tenacious and less likely to make and deliver on commitments. Collapsing conscientiousness is not the only personality shift visible in the data. Neuroticism — a function of the much-discussed increase in anxiety — has risen almost as much. Young adults also report feeling less amiable and outgoing. There’s a particularly steep drop shown in the latter during the pandemic, when young people bore the brunt of restrictions on contact in order to protect others from harm. In fact, long the most extroverted group in society, young adults are now the most introverted.

Takeaways by Macro Roundup® AI

  1. People in their twenties and thirties report being increasingly easily distracted and careless, less tenacious, and less likely to make and deliver on commitments, according to the Understanding America Survey.
  2. Neuroticism, a function of the increase in anxiety, has risen almost as much as conscientiousness has fallen in the Understanding America Survey data.
  3. Young adults, long the most extroverted group in society, now report being the most introverted, with a particularly steep drop during the pandemic, when they bore the brunt of restrictions on contact.

Related Articles:

  • Life Under Two: Debt, Deficits, and the AI Discontinuity — The United States averaged annual GDP growth above 3% for decades, but 2% or less is now more likely. Over ten years, the author estimates that gap amounts to a $3.36tn larger economy, $3.7tn more tax revenue, and 6m more jobs, shifting public psychology from abundance toward scarcity.
  • Zero-Sum Thinking and the Roots of U.S. Political Divides — Zero-sum thinking — viewing gains for one group as losses for another — is associated with greater support for liberal economic policies, stronger alignment with the Democratic Party, and weaker alignment with the Republican Party. The authors also find it linked to believing the QAnon conspiracy theory holds some truth and to sympathy for the attack on the U.S. Capitol.
  • US College Students’ Well-Being — A study of US college students finds moderate-to-severe depression reported by 26% of students at one elite institution, 22% reporting moderate-to-severe anxiety and 10% contemplating suicide. Poorer mental health is associated with being female, little exercise, low grades, and financial aid; financial stress is associated with higher suicidal thoughts.

AI Summary. China's goods exports to the US have fallen sharply—down 22% year-over-year—while exports to the rest of the world grew 12%, almost entirely offsetting the drop. The author argues this shift amounts to a deflationary shock for China, as exporters either transship goods to the US or discount them elsewhere, weighing on profits.

Robin Brooks Robin Brooks Substack
Date Posted:
Is Database:
Database

China’s goods exports to the US are down 22% y/y but have been offset by a 12% increase in goods exports to the rest of the world. Robin Brooks notes that transshipping or selling outside the US will compress exporters’ margins, both acting as deflationary shocks.

Does redirecting exports away from the US create deflation in China?

Core argument: China's exports to the US fell 22% year-over-year in July 2025, while its exports to the rest of the world rose 12% year-over-year, a higher-than-normal growth rate.

The left chart above shows [China's overall goods exports] (black line) and breaks those down into goods going to the US (red bars) and those going to any other country (blue bars). Exports to the US have fallen sharply, with exports to the rest of the world almost completely offsetting this drop. The right chart zeros in on this dynamic. It shows year-over-year growth in exports to the US (red line) versus exports to everywhere else (blue line). In July 2025, China’s exports to the US were down a stunning 22% from a year before, while exports to everywhere else were up 12% year-over-year, a higher-than-normal growth rate. The rise in exports to places other than the US is a deflationary shock for China, which is either transshipping goods to the US (weighing on exporters’ profits) or discounting goods to boost demand for Chinese goods outside the US (also weighing on exporters’ profits).

Takeaways by Macro Roundup® AI

  1. China's exports to the US fell 22% year-over-year in July 2025, while its exports to the rest of the world rose 12% year-over-year, a higher-than-normal growth rate.
  2. China's overall goods exports held up as the sharp drop in shipments to the US was almost completely offset by rising exports to all other countries.
  3. The rise in exports to destinations other than the US amounts to a deflationary shock for China, which is either transshipping goods to the US or discounting goods to boost demand abroad, both weighing on exporters' profits.

Related Articles:

  • China Emerges From Trade Chaos With Record Exports, Surplus — China posted a record half-year trade surplus of about $586bn, with exports up 5.8% year-over-year to $325bn and imports up 1.1%. Shipments to the US fell 16.1% from a year earlier, while exports to the 10 Southeast Asian nations in the Asean group rose 17%.
  • Our Revamped China Wage Tracker Shows Gradual Softening in Wage Growth — A revamped wage tracker estimates China's wage growth slowed to 4.2% year-over-year in the first quarter and 3.9% in the second, down from 4.5% in the prior fourth quarter and about 1 percentage point below official figures. The authors argue sluggish wage growth may weigh on consumption growth.
  • Why US Tariffs Are Deflationary for China — China's exports to the US have fallen sharply while its exports to the rest of the world have risen by an almost offsetting amount. The author argues both adjustments are deflationary for China: routing goods through third countries raises transport costs and squeezes exporter margins, while redirecting goods to new markets requires cutting prices to generate demand.

AI Summary. U.S. flood insurance is provided almost exclusively through the federally managed National Flood Insurance Program, and direct property damage from flooding totaled nearly $144bn while program payouts covered roughly $50bn, or about 35% of that damage. These figures exclude indirect costs such as lost income and production, understating the full economic toll.

Kristian Blickle, Evan Perry, and João Santos Federal Reserve Bank of New York
Date Posted:

Btw 2010 and 2023, direct property damage from flooding totaled $144B (in 2023 $). The National Flood Insurance Program covered 35% of the direct damages.

How much U.S. flood damage does federal flood insurance cover?

Core argument: U.S. direct property damage from flooding totaled nearly $144 billion (2023 USD) between 2010 and 2023, while National Flood Insurance Program payouts on that damage came to roughly $50 billion, just 35% of direct damages.

Flood insurance in the U.S. is almost exclusively provided through the National Flood Insurance Program (NFIP), which is managed by the Federal Emergency Management Agency (FEMA). The U.S. has sustained substantial and largely uninsured damages from flooding over the past fifteen years. The chart displays the cumulative damages from flooding in the U.S. and the cumulative insurance payouts on these damages, starting in 2010. Between 2010 and 2023, the direct property damage from flooding totaled nearly $144 billion (in 2023 USD). In the same period, insurance payments on property damage from the National Flood Insurance Program totaled approximately $50 billion (in 2023 USD)—just 35% of the direct damages. These damage estimates understate the full economic cost of flooding by excluding indirect damages (for example, lost income and production associated with floods).

Takeaways by Macro Roundup® AI

  1. U.S. direct property damage from flooding totaled nearly $144 billion (2023 USD) between 2010 and 2023, while National Flood Insurance Program payouts on that damage came to roughly $50 billion, just 35% of direct damages.
  2. Flood insurance in the U.S. is almost exclusively provided through the National Flood Insurance Program, managed by the Federal Emergency Management Agency.
  3. The $144 billion damage estimate understates the full economic cost of U.S. flooding because it excludes indirect damages such as lost income and production associated with floods.

Related Articles:

  • Insurance: Climbing Costs Cut Into Consumer Budgets — Bank of America internal data show median household property insurance payments (auto and home) rose 6% year-over-year and more than 40% over five years, lifting such payments to 5% of income. Younger generations saw the largest increases, though Gen X payments also rose 45%.
  • Catastrophe Bond Sales Hit Record as Insurers Offload Climate Risks — Insurers are issuing catastrophe bonds, which transfer risk from events such as wildfires, hurricanes and earthquakes to bondholders, at a record pace, with issuance reaching $18.1bn against a prior full-year record of $17.7bn. A Swiss Re index of total returns on these bonds has risen 14% over the past year.
  • The World Is Getting Riskier. Americans Don’t Want to Pay for It — Americans' resistance to paying to insure against risks from illness, old age, and natural and financial disasters is growing alongside those risks; JPMorgan estimated wildfires around Los Angeles caused $50bn in losses, of which only $20bn were insured. Socializing risk removes the price signal that encourages the insured to mitigate their own exposure.

Monday, August 11, 2025

AI Summary. Unemployment has risen by 0.30 percentage points for the fifth of workers most exposed to artificial intelligence, compared with 0.94 percentage points for the least exposed fifth, leaving no visible AI effect on jobs. The share of workers leaving the labor force within a year is lowest among the most AI-exposed workers and roughly flat.

Sarah Eckhardt and Nathan Goldschlag Economic Innovation Group
Date Posted:
Is Database:
Database
Is Important:
Important

The unemployment rate for the quintile of workers most exposed to AI rose 0.30pp between 2022 and early 2025, but for the least-exposed quintile it climbed 0.94pp. “By the most obvious measures, the effect of AI on jobs is invisible.”

Is there any visible effect of AI on employment?

Core argument: Between 2022 and the beginning of 2025, the unemployment rate for the quintile of workers most exposed to AI rose 0.30 percentage points, versus 0.94 percentage points for the least exposed quintile, with the other three quintiles also rising.

Btw 2022 and the beginning of 2025 the unemployment rate for the quintile of the most exposed workers rose by 0.30 percentage points. For the quintile of the least exposed workers, it climbed by 0.94 percentage points. The unemployment rate also climbed during this period for the other three quintiles [Figure 1]. By the most obvious measure, then, the effect of AI on jobs is invisible. But what about less obvious measures? In Figure 2 we widen our lens, showing the percent of workers exiting the labor force by AI-exposure groups. Specifically, for each worker in a given AI-exposure group, we look at the probability that the worker is not in the labor force a year later. Once again the disruptive effect of AI fails to appear. Highly exposed workers aren’t running for the exits and, in fact, the share exiting the labor force is actually the lowest for the most AI-exposed workers, and has been roughly flat since 2022.

Takeaways by Macro Roundup® AI

  1. Between 2022 and the beginning of 2025, the unemployment rate for the quintile of workers most exposed to AI rose 0.30 percentage points, versus 0.94 percentage points for the least exposed quintile, with the other three quintiles also rising.
  2. By the most obvious measure, unemployment by AI-exposure quintile, the effect of AI on jobs is invisible.
  3. The share of workers not in the labor force a year later is lowest for the most AI-exposed workers and has been roughly flat since 2022.

Related Articles:

  • AI Is Wrecking an Already Fragile Job Market for College Graduates — An analysis of labor data finds a shrinking share of each graduating class is landing jobs requiring a bachelor's degree. At the 15 largest tech companies by market value, entry-level hires as a share of total new hires have fallen 50%, and recent graduates made up 7% of new hires, down from 11%.
  • Technological Disruption in the Labor Market — The authors measure US labor market structure since 1880 and find 1990–2017 less disruptive than any earlier period, because past changes from general-purpose technologies like steam and electricity were so profound. They identify four recent signs of accelerating change, including a 50%+ rise in science and technology employment share since 2010.
  • The Kanye/Data Center Crossover — Data center capital spending differs from fiber and railroad investment because most of the cost sits in processors that are replaced every 3-4 years rather than in long-lived shells, power, and cooling connections. That short replacement cycle requires returns high enough to exceed a risk-adjusted cost of capital of 12-14% before the hardware must be replaced.

AI Summary. The US imports roughly 40% of the goods it consumes, directly and through imported inputs, with labor-intensive goods like apparel almost entirely imported. The authors argue America's main comparative disadvantage is its cost of production, driven mainly by labor costs, with real estate costs and a strong dollar also contributing.

Danny DeBois, Aidan Madigan-Curtis, Kate Dunbar and Kenza Wilks Bridgewater Associates
Date Posted:
Is Database:
Database
Is Important:
Important

Bridgewater estimates that closing the ~$3T gap btw US manufacturing output and goods consumption will be an uphill battle. Largely due to manufacturing wages 7x that of China, the “US’s biggest comparative disadvantage is its cost of production.”

What holds back US manufacturing competitiveness against foreign producers?

Core argument: The US imports roughly 40% of the goods it consumes, both directly and indirectly through domestic manufacturers' use of imported inputs.

The US imports roughly 40% of the goods it consumes, both directly and indirectly through domestic manufacturers’ use of imported inputs. Certain goods consumption in the US, like apparel, is almost entirely imported, as labor is a large share of the total cost, and these goods can be produced much more cheaply outside the US. Tariffs on these kinds of goods are especially likely to be passed on to the consumer, given the lack of any viable competition from domestic producers any time soon. The US’s biggest comparative disadvantage is its cost of production. Globalization facilitated the offshoring of manufacturing in the US toward economies with much cheaper costs of production, particularly for goods with relatively low manufacturing value added that are also easy to ship internationally. Though US labor costs are the predominant driver of this cost differential, higher real estate costs also play a role. Energy prices, meanwhile, are a modest support. A secularly strong dollar has also effectively raised the cost of US-produced goods relative to foreign-produced goods and contributed to the decline of US manufacturing competitiveness.

Takeaways by Macro Roundup® AI

  1. The US imports roughly 40% of the goods it consumes, both directly and indirectly through domestic manufacturers' use of imported inputs.
  2. The US's biggest comparative disadvantage in manufacturing is its cost of production, driven predominantly by labor costs with higher real estate costs also contributing, while energy prices offer a modest support.
  3. Tariffs on goods like apparel, where US consumption is almost entirely imported because labor is a large share of total cost, are especially likely to be passed on to consumers given the absence of viable domestic competition.

Related Articles:

  • Why Is Manufacturing Productivity Growth So Low? — Manufacturing productivity growth fell behind other sectors after computer and semiconductor manufacturing productivity, which had grown at roughly 15% and 11% a year, decelerated substantially. The authors also present suggestive evidence that quality improvements are missed in price measures, estimating manufacturing productivity growth is understated by 0.8 percentage points, including 1.7 points for durable goods.
  • The Mysterious Slowdown in U.S. Manufacturing Productivity — US manufacturing labor productivity growth slowed broadly after the Great Recession, with the four largest firms in each industry averaging –0.5% growth from 2010 to 2022 versus 3.7% from 1987 to 2007, and smaller follower firms showing an identical 4.2 percentage-point decline. The authors find the slowdown spans all industries and firm sizes.
  • Help for the Heartland? The Employment and Electoral Effects of the Trump Tariffs in the United States — The authors find that the combined effect of U.S.-China trade war import tariffs, retaliatory tariffs, and farm subsidies on employment in exposed U.S. locations was at best a wash and possibly mildly negative, while residents of tariff-protected areas became less likely to identify as Democrats and more likely to vote Republican.

AI Summary. The authors argue the United States is a decade away from the scale of automated "lights-out" manufacturing needed to meet its own end-to-end demand at comparable cost. With 98% of US manufacturers being small or medium enterprises, low deployment of mature productivity technologies reflects limited commercial adoption and investment rather than technical readiness.

Danny DeBois, Aidan Madigan-Curtis, Kate Dunbar and Kenza Wilks Bridgewater Associates
Date Posted:
Is Database:
Database
Is Important:
Important

98% of manufacturers in the US are small-to-medium enterprises that are relatively underpenetrated by advanced technologies, including robots, additive manufacturing, and embodied AI, posing headwinds to rapid increases in US manufacturing productivity.

What limits automation adoption in US manufacturing?

Core argument: 98% of US manufacturers are small-to-medium enterprises that remain relatively underpenetrated by the technologies capable of driving significant growth in manufacturing productivity.

The United States is a decade away from the scale and scope of “lights-out” manufacturing that could enable it to rival the rest of the world’s manufacturing output to satisfy the US’s own end-to-end demand, at a similar or lesser cost. While the current scale and impact of advanced technologies in American industrial production is low outside of low-volume defense manufacturing, several key technological advances are having a positive impact on US industrial productivity. Unfortunately, 98% of manufacturers in the United States are small-to-medium enterprises and are relatively underpenetrated by technologies capable of driving significant growth in manufacturing productivity. Essentially, the low levels of deployment of these increasingly mature technologies in the US today are limited not by technical readiness or maturity, but by commercial adoption and willingness to invest in advanced capabilities and improved manufacturing throughput.

Takeaways by Macro Roundup® AI

  1. 98% of US manufacturers are small-to-medium enterprises that remain relatively underpenetrated by the technologies capable of driving significant growth in manufacturing productivity.
  2. The United States is a decade away from the scale and scope of "lights-out" manufacturing that could rival the rest of the world's output in satisfying US end-to-end demand at similar or lesser cost.
  3. Low deployment of increasingly mature manufacturing technologies in the US is limited by commercial adoption and willingness to invest in advanced capabilities and improved throughput, not by technical readiness or maturity.

Related Articles:

  • America Is Missing The New Labor Economy – Robotics Part 1 — Building an identical robotic arm in the US costs about 2.2x more than in China, and US-labeled components still depend heavily on Chinese parts and materials with no scalable alternative. The author argues China is the only country positioned to capture full-scale intelligent robotic automation, which would let it outcompete the US across production.
  • Global Robot Density in Factories Doubled in Seven Years — Global factory robot density reached a record 162 robots per 10,000 manufacturing employees, more than double the level of seven years earlier. The European Union averages 219 units per 10,000 employees, North America 197, and Asia 182, with Korea, Singapore, mainland China and Japan among the ten most automated economies.
  • How AI & Robots are Smashing Economics, and Why It Matters — As artificial intelligence and robots make wages less responsive to employment conditions, the link between tight labor markets and inflation weakens, undermining central banks' inflation-targeting frameworks. The author argues central banks would then need far larger interest-rate moves to achieve the same effect, producing more volatile and less predictable outcomes.

AI Summary. Interest rate futures price about 130 basis points of Federal Reserve rate cuts through the end of next year, versus no cuts expected from the European Central Bank and 50 and 70 basis points for the Bank of England and Reserve Bank of Australia. Possible explanations include tariff-driven US recession risk, the Fed's slower easing so far, and political pressure.

Robin Brooks Robin Brooks Substack
Date Posted:
Is Database:
Database

Markets have priced in five 25bp cuts for the Fed but no cuts for the ECB through the end of 2026. This may reflect the 200bp ECB cuts that have already occurred or concerns about a US slowdown.

Do expected interest rate cuts explain the dollar's decline?

Core argument: Interest rate futures priced 130 basis points of Fed cuts through the end of next year, slightly more than five 25 basis point cuts, versus an ECB expected to stay on hold.

[The chart] uses interest rate futures to calculate how much markets expect key central banks to cut their policy rates through the end of next year. As of Friday, markets priced 130 basis points in cuts for the Fed (so slightly more than five 25 basis point cuts) in contrast to an ECB that’s expected to be on hold. Only the Bank of England (GB) and Reserve Bank of Australia (AU) come close in the eyes of the market, with 50 and 70 basis points priced, respectively. Why are market pricing so many more cuts for the Fed? I can think of three possible reasons. The first is that they think tariffs will boomerang on the US, forcing the Fed to cut aggressively as recession risk mounts and holds down any inflation impulse from tariffs. The second is that the Fed last cut in December of last year while other central banks kept cutting. For example, the ECB has already made four 25bp cuts this year, bringing this easing cycle to 200 basis points, double that of the Fed. The third is that markets price aggressive Fed easing not because of economic fundamentals but because of mounting political pressure.

Takeaways by Macro Roundup® AI

  1. Interest rate futures priced 130 basis points of Fed cuts through the end of next year, slightly more than five 25 basis point cuts, versus an ECB expected to stay on hold.
  2. The Bank of England and Reserve Bank of Australia come closest to the Fed in market-priced easing, with 50 and 70 basis points of cuts respectively.
  3. Three possible explanations for the Fed's larger priced easing are tariffs boomeranging on the US as recession risk mounts, the Fed's pause since last December while the ECB cut 200 basis points this cycle, and mounting political pressure.

Related Articles:

  • Explaining the US Dollar’s Depreciation in the First Half of 2025 — The dollar depreciated by more than interest-rate differentials alone would predict, and regression analysis attributes the gap to the trade war and economic policy uncertainty, including foreign investors' concerns about proposed taxes on foreign holders of US assets and a possible currency accord. The author expects the dollar to appreciate as these uncertainties resolve.
  • Dollar Positioning — The dollar is having its worst year since 1973, and the author argues that persistent large fiscal deficits and challenges to Federal Reserve independence imply significant further downside for a currency still trading above historical averages. Foreign investors hold very large dollar asset exposure, so a shift from currency hedging to outright selling could spur a self-reinforcing cycle of declines.
  • A Trump Risk Premium in the Dollar — Movements in the trade-weighted US dollar against other major currencies map almost entirely onto changes in forward interest-rate differentials, with the same positive relationship holding across periods despite noisier data. The author argues this leaves no discernible evidence of a separate political risk premium in the dollar.

AI Summary. The US government negotiated an arrangement taking 15% of Nvidia's Chinese sales of its H20 artificial intelligence processor, a chip with performance restricted relative to versions sold to US customers. The administration said it would consider allowing Nvidia to sell China more advanced chips based on its newer Blackwell platform.

Michael Acton, Demetri Sevastopulo, Tim Bradshaw and Myles McCormick Financial Times
Date Posted:

In exchange for export licences to export H20 chips to China which had been previously banned by US export controls, Nvidia, as well as AMD, will turn 15% of their revenue from China sales over to the US government.

Will the US permit advanced AI chip sales to China?

Core argument: Trump confirmed an unprecedented deal with Nvidia CEO Jensen Huang giving the US government 15% of Nvidia's Chinese sales of its H20 processor.

Donald Trump said he will consider allowing Nvidia sell a more advanced artificial intelligence chip in China, after confirming he “negotiated a little deal” to give Washington a share of the chipmaker’s revenues from the country. The unprecedented deal Trump struck with Nvidia’s chief executive Jensen Huang [involves] the US taking 15% of Chinese sales of its H20 processor, which it introduced in 2023 to comply with Biden-era controls on AI chips. The performance of the H20 chips sold to China is restricted compared with those available to US customers. The H20 is based on Nvidia’s older Hopper architecture but Trump on Monday said that he planned to discuss a new deal with Huang to allow it to sell chips to China based on its latest Blackwell platform.

Takeaways by Macro Roundup® AI

  1. Trump confirmed an unprecedented deal with Nvidia CEO Jensen Huang giving the US government 15% of Nvidia's Chinese sales of its H20 processor.
  2. Nvidia introduced the H20 in 2023 to comply with Biden-era AI chip controls, and its performance is restricted relative to chips sold to US customers.
  3. Trump said he plans to discuss a new deal with Huang to allow Nvidia to sell China chips based on its latest Blackwell platform, rather than the older Hopper architecture underlying the H20.

Related Articles:

  • 2025 AI Diffusion Export Controls – Microsoft Regulatory Capture, Oracle Tears, Impacts Quantified, Model Restrictions — The US Framework for Artificial Intelligence Diffusion imposes export restrictions aimed at preventing China from accessing the computing power needed to build frontier AI models. It sorts the world into three tiers — the US and 18 allies, 23 arms-embargoed nations plus Macau, and all other countries — with differing access to AI chips and caps on graphics processor purchases.
  • China’s AI Industry Has Almost Caught Up With America’s — Chinese developer DeepSeek trained its V3 large language model using fewer than 3m chip-hours at an estimated cost under $6m—roughly a tenth of the computing power and expense of Meta's Llama 3.1—through a series of marginal efficiency improvements rather than one breakthrough. The model was trained on chips constrained by American export sanctions.

AI Summary. China is seeking relief from US semiconductor export controls as part of a trade deal, and people familiar with the matter say Beijing is more concerned about restrictions on high-bandwidth memory chips than about the H20 chip Nvidia designed for the Chinese market. Those memory controls constrain Chinese companies' ability to develop their own AI chips.

Demetri Sevastopulo Financial Times
Date Posted:

In exchange for a trade deal, China wants American export controls on high-bandwidth memory (HBM) chips, key components for AI chips, relaxed. The chips would allow Huawei to manufacture AI chips that are more competitive with Nvidia’s.

Which chip export controls does Beijing most want lifted?

Core argument: China is pressing the US to relax chip export controls as part of a trade deal, after Biden in 2022 unveiled measures to damage Chinese efforts to buy or make advanced AI chips.

Beijing has been frustrated at US export controls since Biden in 2022 unveiled measures to damage Chinese efforts to buy or make advanced AI chips. In 2024, he banned HBM exports to China to hamper Huawei and Semiconductor Manufacturing International Corporation (SMIC), a Chinese chipmaker. The recent focus on export controls to China has been on the H20 chip Nvidia designed for the Chinese market after Biden banned the export of more advanced chips. The US approved H20 export licences on Friday after Nvidia CEO Jensen Huang met Trump. But people familiar with the matter said China is much more worried about HBM controls because they seriously constrain the ability of Chinese companies, including Huawei, to develop their own AI chips.

Takeaways by Macro Roundup® AI

  1. China is pressing the US to relax chip export controls as part of a trade deal, after Biden in 2022 unveiled measures to damage Chinese efforts to buy or make advanced AI chips.
  2. People familiar with the matter say China is far more worried about the 2024 ban on HBM exports, which seriously constrains the ability of Chinese companies including Huawei to develop their own AI chips.
  3. The US approved export licences for Nvidia's H20 chip, designed for the Chinese market after Biden banned exports of more advanced chips, following a meeting between Nvidia CEO Jensen Huang and Trump.

Related Articles:

  • Donald Trump Freezes Export Controls To Secure Trade Deal With China — The US Bureau of Industry and Security, which administers export controls, has been instructed to avoid tough measures against China, according to current and former US officials. Officials cite the risk that China retaliates by restricting exports of critical rare earths and magnets to the US, as it has previously done.
  • 2025 AI Diffusion Export Controls – Microsoft Regulatory Capture, Oracle Tears, Impacts Quantified, Model Restrictions — The US Framework for Artificial Intelligence Diffusion imposes export restrictions aimed at preventing China from accessing the computing power needed to build frontier AI models. It sorts the world into three tiers — the US and 18 allies, 23 arms-embargoed nations plus Macau, and all other countries — with differing access to AI chips and caps on graphics processor purchases.
  • Chinese Companies Resort To Repurposing Nvidia Gaming Chips For AI — Chinese firms are developing artificial intelligence tools using processors stripped from Nvidia consumer gaming graphics cards and remounted on new circuit boards, after US export controls blocked sales of high-performance chips. Industry experts describe the practice as a rough workaround for China's shortage of high-end processors.