Edward Conard

Top Ten New York Times Bestselling Author

  • “A full-throated defense of economic dynamism.” - The Wall Street Journal
  • “Unintended Consequences should be read by anyone who takes for granted the superiority of progressive taxation and has not thought carefully about the trade-offs involved.” - The New Republic
  • “…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
  • “…a fresh argument for the productive value of inequality.” - David Autor, Professor of Economics, Massachusetts Institute of Technology
  • “…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
  • “There are an amazing number of good ideas and interesting points made in Unintended Consequences. The thinking underlying it, and the obvious depth of understanding of the author, are very impressive.” - Steven Levitt, coauthor of Freakonomics; 2004 John Bates Clark Medal
  • “…a very valuable contribution.” - Larry Summers, former Secretary of the Treasury and director of the National Economic Council, president emeritus, Harvard University
  • “…a comprehensive explanation of the modern economy.” - Julian Robertson, Founder, Tiger Management
  • “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
  • “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
Upside of Inequality Oxford Unintended Consequences

Macro Roundup


Thursday, February 6, 2025

AI Summary. Estimates of workplace AI adoption vary widely with survey design: the U.S. Census Bureau's biweekly Business Trends and Outlook Survey puts firm-level adoption at 5%, while employment-weighted firm surveys and individual worker surveys range from about 20% to 40%. All available surveys show adoption growing rapidly.

Leland Crane, Michael Green, and Paul Soto Federal Reserve Board
Date Posted:

Crane et al find that existing studies of AI adoption by firms and workers show that between 20-40% of workers are using AI in the workplace, with much higher rates in occupations like computer programming. All surveys suggest rapid growth in adoption.

Why do estimates of workplace AI adoption vary so widely?

Core argument: The U.S. Census Bureau's Business Trends and Outlook Survey, conducted every two weeks, puts firm-level AI adoption at 5 percent, below other surveys, with employment weighting and question phrasing closing much of the gap.

Firm-level adoption may be measured as quite low or fairly high depending on the details of the question asked and the weighting used. Notably, only a few of these surveys were conducted on a recurring basis. A standout example is the top-line U.S. Census Bureau's Business Trends and Outlook Survey (BTOS), which is conducted every two weeks, and shows a number of 5 percent, much lower than other surveys. Employment weighting and the phrasing of the question closes much of the gap. Treating the other firm-level surveys as approximately employment-weighted, estimates of AI adoption range from about 20% to 40%. Individual level surveys of AI use at work find a similar range of adoption rates. Similar levels of individual AI adoption and employment-weighted firm AI adoption can be mutually consistent if some workers are using AI at work without the knowledge of their managers. Regardless of measured AI adoption levels, all available surveys show adoption growing rapidly.

Takeaways by Macro Roundup® AI

  1. The U.S. Census Bureau's Business Trends and Outlook Survey, conducted every two weeks, puts firm-level AI adoption at 5 percent, below other surveys, with employment weighting and question phrasing closing much of the gap.
  2. Treating other firm-level surveys as approximately employment-weighted places AI adoption at roughly 20 to 40 percent, a range similar to that found in individual-level surveys of AI use at work.
  3. Similar rates of individual AI use and employment-weighted firm adoption can be mutually consistent if some workers use AI at work without their managers' knowledge, and all available surveys show adoption growing rapidly.

Related Articles:

  • The Rapid Adoption of Generative AI — A nationally representative U.S. survey found 39% of people aged 18-64 used generative AI, with more than 24% of workers using it in the prior week and nearly one in nine using it every workday. The authors estimate 0.5-3.5% of U.S. work hours are currently assisted by generative AI.
  • 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.
  • A Data-Driven Case That AI Has Already Changed The U.S. Labor Market — Retail employment began declining in the early 2010s after holding steady for more than two decades, coinciding with heavy big tech investment in deep learning and the commercial adoption of AI-directed warehouse robots and predictive pricing and inventory tools. The author argues this circumstantial evidence suggests AI has already affected the US labor market, but does not constitute proof.

AI Summary. US student test scores peaked in the early 2010s; from that peak to the pandemic, scores fell for students at the 10th and 25th percentiles while rising slightly at the 75th and 90th. After the pandemic, scores fell across all percentile groups, with much larger declines among lower-performing students.

Nat Malkus American Enterprise Institute
Date Posted:

.@natmalkus notes the decline in US fourth-grade math scores since 2013 has been driven by large declines at the bottom of the distribution, while the scores of top 10% performers have been relatively stable.

Are declines in US student test scores concentrated among lower-performing students?

Core argument: US student test scores across numerous assessments peaked in the early 2010s, and indexing each percentile group's score to its 2013 peak shrinks the figure's scale from 100 to 30 points.

Student test scores across numerous assessments peaked in the early 2010s. In Figure 3, Panel B, I show how percentile scores change from their peak in 2013. Subtracting the peak 2013 score for each group from that group’s score in a given year makes the 2013 score for each group equal to zero and shrinks the figure’s scale markedly, from 100 to 30 points. Panel B makes the relative changes within percentile groups come into sharp relief: between 2013 and the pandemic, scores for the 10th and 25th percentiles fell, while those for the 75th and 90th percentiles rose slightly. After the pandemic, scores fell across the board, but much more for students in the lower percentiles.

Takeaways by Macro Roundup® AI

  1. US student test scores across numerous assessments peaked in the early 2010s, and indexing each percentile group's score to its 2013 peak shrinks the figure's scale from 100 to 30 points.
  2. Between 2013 and the pandemic, scores for students at the 10th and 25th percentiles fell while scores at the 75th and 90th percentiles rose slightly.
  3. After the pandemic, US student scores fell across all percentile groups, with much larger declines for students in the lower percentiles.

Related Articles:

  • Do Adults Have the Skills They Need to Thrive in a Changing World? — An international assessment of adult skills describes numeracy proficiency levels ranging from counting objects and adding small whole numbers at the lowest level, to extracting information from tables and graphs and performing basic arithmetic with money and percentages at Level 1, up to analysing complex mathematical information and critically evaluating whether data support a claim at Level 5.
  • TIMSS Shows the Bottom Is Falling Out for US Test Scores — US math scores on the international Trends in Mathematics and Science Study fell 18 points for fourth graders and 27 points for eighth graders, on a scale where about 30 points equates to a year of learning. Declines were concentrated among the lowest-performing students, with fourth-grade scores above the 75th percentile barely moving.
  • 2024 SAT Data Drop — Race gaps in SAT scores are consistent across years, with most differences within the range expected from sampling error; the exceptions are Amerindian scores falling and Asian scores rising relative to White scores. The author attributes those shifts to changes in who selects into SAT test-taking rather than genuine population changes.

AI Summary. The US Treasury Secretary said the administration's effort to lower borrowing costs targets 10-year Treasury yields rather than the Federal Reserve's short-term benchmark rate, arguing that spending cuts, a smaller government, and lower energy prices can deliver non-inflationary growth. The Treasury maintained its guidance to keep sales of longer-term debt unchanged.

Viktoria Dendrinou Bloomberg
Date Posted:

Scott Bessent reported that he and President Trump are “focused on the 10-year Treasury [yield].” When asked if Trump wants lower rates, Bessent noted, “[Trump] is not calling for the Fed to lower rates.”

Can spending cuts and lower energy prices bring down long-term yields?

Core argument: Treasury Secretary Scott Bessent said the Trump administration's focus for lowering borrowing costs is the 10-year Treasury yield rather than the Federal Reserve's benchmark short-term rate.

Treasury Secretary Scott Bessent said the Trump administration’s focus with regard to bringing down borrowing costs is 10-year Treasury yields, rather than the Federal Reserve’s benchmark short-term interest rate. “The bond market is recognizing that” under Trump “energy prices will be lower and we can have non-inflationary growth,” Bessent said of the drop in yields in recent weeks. “We cut the spending, we cut the size of government, we get more efficiency in government. And we’re going to go into a good interest-rate cycle. ”Earlier, the US Treasury maintained its guidance on keeping sales of longer-term debt unchanged well into 2025, despite Bessent having criticized the issuance strategy of his predecessor before he was picked for the job.

Takeaways by Macro Roundup® AI

  1. Treasury Secretary Scott Bessent said the Trump administration's focus for lowering borrowing costs is the 10-year Treasury yield rather than the Federal Reserve's benchmark short-term rate.
  2. Bessent attributed the recent drop in yields to the bond market recognizing that energy prices will be lower under Trump and that non-inflationary growth is possible, pointing to cuts in spending and the size of government.
  3. The US Treasury maintained its guidance to keep sales of longer-term debt unchanged well into 2025, despite Bessent having criticized his predecessor's issuance strategy before being picked for the job.

Related Articles:

  • Data Update 4 for 2025: Interest Rates, Inflation and Central Banks! — Interest rates are ultimately driven by macroeconomic fundamentals, with an "intrinsic risk-free rate" defined as the sum of the inflation rate and real GDP growth tracking the observed 10-year Treasury bond yield closely. On this view, central banks' influence comes from preferential access to data on those fundamentals, their capacity to alter them, and their credibility.
  • ATI: Activist Treasury Issuance and the Tug-of-War Over Monetary Policy — The authors argue that by shifting its debt issuance toward shorter maturities, the US Treasury is managing financial conditions in a way that takes on core Federal Reserve functions, a practice they call "activist Treasury issuance." They estimate it has lowered 10-year yields by about 0.25 percentage points, comparable to a one-point cut in the Federal Reserve's policy rate.
  • Resilience Redux in the US Treasury Market — The authors argue that large, persistent US federal deficits and limited dealer balance-sheet flexibility will cause outstanding Treasury debt to grow rapidly relative to the market's capacity to intermediate trades, absent significant changes in market structure. They contend broad central clearing, all-to-all trading, real-time post-trade reporting, capital-regulation changes, and transparent official purchase programs could strengthen market capacity and resilience.

AI Summary. Analysts estimate that enacting $5.5tn of unoffset tax cuts and spending would raise federal interest costs by about $1.3tn over a decade, lifting annual net interest payments to $2tn, or 4.7% of GDP, by 2035 — above the prior record of 3.2% — and interest as a share of revenue to 27%.

CRFB Staff Committee for a Responsible Federal Budget
Date Posted:
Is Database:
Database

The House preliminary reconciliation bill would allow up to $5.5T of net primary deficit. @MarcGoldwein estimates that if passed, the bill would increase interest costs by ~ $1.3T over the decade, rising to $2T or 4.7% of GDP by 2035.

How much would unoffset tax cuts raise federal interest costs?

Core argument: The authors estimate that $5.5 trillion of unoffset new and extended tax cuts and spending would raise federal interest costs by about $1.3 trillion over the next decade, exceeding $250 billion per year by 2035.

We estimate that passing an additional $5.5 trillion of new and extended tax cuts and spending without offsets would: boost interest costs by about $1.3T over the next decade, exceeding $250 billion per year by 2035. Increase annual federal net interest payments to $2T or 4.7% of GDP by 2035, far above the previous record of 3.2%. Boost the interest-to-revenue ratio from 9% back in 2021 and 18% today to 27% by 2035 (compared to 22% under current law). Increase the risk of a debt spiral by pushing up long-term interest rates (R) and potentially slowing the long-term growth rate (G).

Takeaways by Macro Roundup® AI

  1. The authors estimate that $5.5 trillion of unoffset new and extended tax cuts and spending would raise federal interest costs by about $1.3 trillion over the next decade, exceeding $250 billion per year by 2035.
  2. Annual federal net interest payments would reach $2 trillion, or 4.7% of GDP, by 2035 under the $5.5 trillion package, above the previous record of 3.2%, according to the authors' estimates.
  3. The interest-to-revenue ratio would climb from 9% in 2021 and 18% today to 27% by 2035 under the package, compared with 22% under current law, with the authors warning of a heightened debt-spiral risk as long-term interest rates rise and growth potentially slows.

Related Articles:

  • The Budget and Economic Outlook: 2025 to 2035 — The Congressional Budget Office projects the federal deficit will grow from $1.9tn (6.2% of GDP) to $2.7tn by 2035, well above the 3.8% of GDP average of the past 50 years. Outlays are projected to rise to 24.4% of GDP, driven by growth in Social Security and Medicare spending and rising net interest costs, while revenues reach 18.3%.
  • Debt-to-GDP Growth Could Double with a $5.5 Trillion Bill — A budget package allowing up to $5.5tn of net deficit increases, largely through tax cuts, is estimated to add $6.5tn to the debt with interest over ten years, pushing debt from 100% of GDP to 133%. That 33 percentage-point rise is nearly double the 18-point increase projected under current law.
  • Long Rates Are Too High — A simple model of the relationship between long-term interest rates and Federal Reserve policy expectations shows 10-year rates sitting 40 basis points above what those expectations would predict, consistent with increases in term premium measures from the New York and San Francisco Federal Reserve Banks. Markets worry this extra premium reflects fears about fiscal sustainability.

AI Summary. Asteroid 2024 YR4, detected by a NASA-funded network of near-Earth object telescopes, is estimated by the European Space Agency to be between 130 and 330 feet long, a range derived from the sunlight it reflects because its surface reflectivity is unknown. A more precise radar measurement will be possible when it makes another safe close pass by Earth.

Robin George Andrews New York Times
Date Posted:

A newly discovered asteroid, estimated at 130-330 ft. long, has a 2.3% chance of hitting Earth in 2032 according to NASA. Last year NASA’s Double Asteroid Redirection Test (DART) changed the orbit of a large asteroid in a defense experiment..

How precisely can astronomers measure a newly detected asteroid's size?

Core argument: Asteroid 2024 YR4 measures between 130 and 330 feet long, according to the European Space Agency's Near-Earth Object Coordination Centre, a range derived from the sunlight it reflects because its surface reflectivity is unknown.

Asteroid 2024 YR4 was identified by the Asteroid Terrestrial-impact Last Alert System four telescopes around the world that hunt near-Earth objects and are funded by NASA. Its telescope in Chile found 2024 YR4 on Dec. 27, just two days after a close approach to Earth. It’s now speeding away from the planet and getting fainter by the day. According to the European Space Agency’s Near-Earth Object Coordination Centre, it is between 130 and 330 feet long. This size range is based on the amount of sunlight it is reflecting. Without knowing exactly how reflective 2024 YR4’s surface is, only a range of sizes can be given. A more accurate estimate could be made using radar, but that won’t be possible until the asteroid makes another close, but perfectly safe, pass by Earth on Dec. 17, 2028. A 130-foot asteroid is comparable to the Tunguska impactor, a meteor that exploded over a remote patch of Siberia in 1908 and decimated an 800-square-mile forest.

Takeaways by Macro Roundup® AI

  1. Asteroid 2024 YR4 measures between 130 and 330 feet long, according to the European Space Agency's Near-Earth Object Coordination Centre, a range derived from the sunlight it reflects because its surface reflectivity is unknown.
  2. The NASA-funded Asteroid Terrestrial-impact Last Alert System's telescope in Chile found 2024 YR4 on Dec. 27, two days after the asteroid's close approach to Earth, and it is now speeding away and growing fainter.
  3. A more accurate size estimate of 2024 YR4 using radar must wait until the asteroid makes another close but safe pass by Earth on Dec. 17, 2028.

Related Articles:

  • It’s Not a Space Race, It’s the Space Olympics — The US operates 81% of commercial satellites in orbit and 86% of global effective launch capacity, with NATO allies second in satellites at 11% and China second in launch capacity at 9%. China is pursuing mega satellite constellations and reusable rockets, in many cases copying approaches already proven by US companies.
  • Building an Enduring Advantage in the Third Space Age — The United States and China each logged their highest-ever annual launch rates, at 103 and 66 launches, while Russia fell to 19 and Europe recorded three. The US accounted for roughly half of global launches but 81% of effective launch capacity, since Falcon 9 and Falcon Heavy comprised 93% of US launches.

Friday, February 7, 2025

AI Summary. Estimates put algorithmic progress in artificial intelligence at roughly 4x per year, meaning the compute needed for a given capability level falls 4x annually, with some industry leaders arguing improvements run closer to 10x. The Bank of China announced a $140bn (1tn yuan) five-year subsidy for the AI industry chain aimed at making China self-reliant in science and technology.

Dylan Patel, AJ Kourabi, Doug O'Laughlin, and Reyk Knuhtsen Semianalysis
Date Posted:

.@dylan522p @SemiAnalysis_ notes that DeepSeek’s efficiency gains are accelerating the decline in AI inference costs, tracking the trend from GPT-4 to DeepSeek-V3 and notes, “By the end of the year do not be shocked if costs fall another 5x.”

How fast is the compute needed for given AI capability falling?

Core argument: Estimates put AI algorithmic progress at 4x per year, meaning 4x less compute is needed annually for the same capability, while Anthropic CEO Dario argues advances run faster, yielding up to 10x improvement.

Estimates put algorithmic progress at 4x per year, meaning that for every passing year, 4x less compute is needed to achieve the same capability. Dario, CEO of Anthropic argues that algorithmic advancements are even faster and can yield a 10x improvement. When investigating the cost for GPT-4, in this case, we see algorithmic improvements and optimizations creating a 10x decrease in cost and increase in capability. To be clear DeepSeek is unique in that they achieved this level of cost and capabilities first. They are unique in having released open weights, but prior Mistral and Llama models have done this in the past too. DeepSeek has achieved this level of cost but by the end of the year do not be shocked if costs fall another 5x. The Bank of China has announced a $140B (1T Yuan) subsidy over the next 5 years for the AI Industry Chain, one day after meeting the CEO and founder of DeepSeek. The goal of this subsidy is explicitly to empower China to be completely self-reliant in science and technology.

Takeaways by Macro Roundup® AI

  1. Estimates put AI algorithmic progress at 4x per year, meaning 4x less compute is needed annually for the same capability, while Anthropic CEO Dario argues advances run faster, yielding up to 10x improvement.
  2. GPT-4's cost shows algorithmic improvements and optimizations producing a 10x decrease in cost alongside increased capability, with DeepSeek first to reach this cost-capability level, though Mistral and Llama models also released open weights previously.
  3. The Bank of China announced a $140 billion (1T yuan) subsidy for the AI industry chain over five years, one day after meeting DeepSeek's CEO and founder, explicitly to make China completely self-reliant in science and technology.

Related Articles:

  • DeepSeek and the Sincerest Form of Flattery — DeepSeek appears to have trained its models about 45x more efficiently than other leading-edge models, combining existing techniques—activating only 37bn of 671bn parameters at a time, a smaller memory cache, lower-precision training, and multi-token prediction—alongside self-reinforcement learning. Training cost $5.6m, and its interface charges $0.14 per million tokens versus OpenAI's $7.50.
  • DeepSeek FAQ — Consumers and businesses stand to gain most from a future of effectively free AI products, along with big consumer tech companies, since a world of free AI makes product and distribution the decisive advantages they already hold. China also benefits, and the United States faces a choice between expanding chip export restrictions and permission-based regulation or allowing itself to compete.
  • 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.

AI Summary. Microsoft, Alphabet, Amazon and Meta reported combined capital expenditure of $246bn, up from $151bn the prior year, and forecast spending could exceed $320bn as they compete to build data centres filled with specialised AI chips. Capital spending among the "Magnificent Seven" rose 40%, versus 3.5% for the other 493 S&P 500 companies.

Stephen Morris and Rafe Uddin Financial Times
Date Posted:
Is Database:
Database

The capex of Meta, Google, Microsoft, and Amazon is forecast to grow to at least $320B this year, up 30% y/y, “to remain at the forefront of AI large language model research.”

How much are big tech firms spending to build AI capacity?

Core argument: Microsoft, Alphabet, Amazon and Meta reported combined capital expenditure of $246bn in 2024, up from $151bn in 2023, and forecast spending could exceed $320bn this year.

Microsoft, Alphabet, Amazon and Meta have reported combined capital expenditure of $246bn in 2024, up from $151bn in 2023. They forecast spending could exceed $320bn this year as they compete to build data centres and fill them with clusters of specialised chips to remain at the forefront of AI large language model research. Spending among the “Magnificent Seven” — which also includes Apple, Nvidia and Tesla — dwarfs the rest of the US benchmark S&P 500. Their capital spending rose 40% in 2024 compared with 3.5% among the remaining 493 companies, according to Société Générale. Profits among the elite group soared by a third in the same period, versus 5% among the rest.

Takeaways by Macro Roundup® AI

  1. Microsoft, Alphabet, Amazon and Meta reported combined capital expenditure of $246bn in 2024, up from $151bn in 2023, and forecast spending could exceed $320bn this year.
  2. Microsoft, Alphabet, Amazon and Meta are competing to build data centres and fill them with clusters of specialised chips to remain at the forefront of AI large language model research.
  3. Capital spending among the "Magnificent Seven" rose 40% in 2024 versus 3.5% among the other 493 S&P 500 companies, while their profits rose by a third versus 5%, according to Société Générale.

Related Articles:

  • DeepSeek Raises Micro Risks, Macro Upside — More efficient AI model training and falling compute costs pose a near-term risk of lower AI-related capital spending, which equity analysts project will rise to $325bn by the end of 2025, though that risk is seen as limited. Cheaper, more competitive models could instead speed the buildout of AI platforms and applications, raising macroeconomic upside.
  • A Severe Case of COVIDIA: Prognosis For An AI-Driven US Equity Market — Computing cycles move from infrastructure to platforms to applications, and NVIDIA data center revenues as a share of market-wide capital spending are projected to reach levels last seen at the 1969 mainframe peak and the dot-com boom. Large cloud providers spend more than twice as much training foundational models as running them for customers.
  • Gen AI: Too Much Spend, Too Little Benefit? — The author argues generative AI is unlikely to drive revenue expansion because it helps employees find information faster rather than find better information. If AI's benefits remain limited to efficiency gains, competitors can adopt the same tools, so cost savings get arbitraged away rather than raising prices or margins.

AI Summary. European stocks have reached record highs but trail US equities, with the Stoxx Europe 600 posting its widest dollar-terms underperformance versus the S&P 500 since the index began in 1998, and US stocks trading at 22x earnings, a near 60% premium over Europe. When companies follow capital abroad, headquarters, management, and hiring can move with it.

Sagarika Jaisinghani, Pablo Mayo Cerqueiro, Jan-Patrick Barnert, and Neil Campling Bloomberg
Date Posted:
Is Database:
Database

“US stocks on average trade at 22 times earnings, a near 60% premium over Europe.” As EU firms choose to list in the US, “headquarters, management and hiring can be pulled away too” and Europe “loses out on the next generation of ideas and companies.”

Do fragmented capital markets hold back European equities and companies?

Core argument: The Stoxx Europe 600 in 2024 posted its biggest dollar-terms underperformance versus the S&P 500 since the European benchmark's 1998 creation, even as a global rally pushed the index to record highs.

Although a global rally has lifted the Stoxx Europe 600 Index to record highs, its performance pales in comparison to the US. In 2024, the benchmark suffered its biggest underperformance in dollar terms versus the S&P 500 since the regional benchmark was created in 1998. US stocks on average trade at 22 times earnings, a near 60% premium over Europe. Scare stories about capital markets aren’t just about share prices or investor returns. When companies are driven to follow the money, then headquarters, management and hiring can be pulled away too. If they don’t have enough incentive to stay, the continent loses out on the next generation of ideas and companies it needs to stay relevant.

Takeaways by Macro Roundup® AI

  1. The Stoxx Europe 600 in 2024 posted its biggest dollar-terms underperformance versus the S&P 500 since the European benchmark's 1998 creation, even as a global rally pushed the index to record highs.
  2. US stocks trade at an average 22 times earnings, a near 60% premium over European equities.
  3. When companies are driven to follow the money, headquarters, management and hiring can be pulled away from Europe, costing the continent the next generation of ideas and companies it needs to stay relevant.

Related Articles:

  • Why the U.S. Economy Is Trouncing Europe’s — Only about 12% of Americans score at the highest levels on international academic tests versus 34% at the lowest, yet the U.S. has created nearly 250 companies worth over $10bn ($30tn in value) against Europe's 14, and median U.S. disposable household income is 25% above Germany's. The author argues expanding high-skilled immigration is the fastest way to accelerate growth.
  • The Future of European Competitiveness – A Competitiveness Strategy for Europe — A European Union competitiveness report attributes the EU's labour productivity growth gap with the US largely to weaker activity in information and communications technology and large-scale digital services; excluding main ICT sectors, EU productivity was broadly at par with the US from 2000-2019. It estimates Europe's investment share must rise about 5 percentage points of GDP.
  • A Visualization of Europe’s Non-Bubbly Economy — The Draghi report notes that no European Union company with a market capitalization above €100bn was founded from scratch in the past fifty years, while all six US companies valued above €1tn were. The 13 EU companies founded in that period and worth at least $10bn total about $400bn, versus nearly $30tn for their US counterparts.

AI Summary. The paper proposes making the high-tax exclusion mandatory under the US minimum tax on foreign corporate earnings, with the exclusion threshold set at the minimum tax's effective rate, so the tax applies only to income below that threshold. The authors estimate this would raise $68.5bn–$72.9bn over a decade while reducing tax-avoidance opportunities.

Kyle Pomerleau American Enterprise Institute
Date Posted:

.@kpomerleau argues that mandatory exclusion of high-tax foreign income for Global Intangible Low-Tax Income (GILTI) calculations would “reduce incentives to shift mobile assets to foreign jurisdictions” and raise $68.5-72.9B in 2025-35.@kpomerleau

Would a mandatory high-tax exclusion raise revenue and curb tax avoidance?

Core argument: Making the GILTI high-tax exclusion mandatory would raise between $68.5 billion and $72.9 billion between 2026 and 2035, according to the paper.

An area that deserves close inspection is Global Intangible Low-Tax Income’s structure as an overall minimum tax on foreign operations. This paper outlines a straightforward reform to GILTI: a mandatory high-tax exclusion. Under current law, corporations can elect to exclude the profits and related foreign taxes that face an effective tax rate [ETR} above a certain threshold. This reform would make this election mandatory and, as a result, GILTI would only apply to income below the exclusion threshold. The effective rate test for the mandatory high-tax exclusion, which is currently 18.9 percent (90 percent of the statutory tax rate of 21 percent), should be set at GILTI’s ETR. For example, if the corporate tax rate is 21 percent and the GILTI ETR is 13.125 percent, the high tax exclusion ETR test should be set at 62.5 percent of the domestic tax rate or 13.125 percent. This reform would reduce opportunities for tax avoidance and reduce incentives to shift mobile assets to foreign jurisdictions while having a minimal impact on complexity, [and] would raise btw $68.5 and $72.9 billion btw 2026 and 2035.

Takeaways by Macro Roundup® AI

  1. Making the GILTI high-tax exclusion mandatory would raise between $68.5 billion and $72.9 billion between 2026 and 2035, according to the paper.
  2. The paper proposes setting the high-tax exclusion effective rate test at GILTI's own effective tax rate — 13.125 percent, or 62.5 percent of a 21 percent corporate rate — rather than the current 18.9 percent threshold.
  3. Under current law corporations may elect to exclude foreign profits and related foreign taxes facing an effective tax rate above a threshold; making that election mandatory would apply GILTI only to income below the exclusion threshold.

Related Articles:

  • Making the Tax Cuts and Jobs Act Permanent: Two Revenue-Neutral, Pro-Growth Options for Tax Reform — Following the 2017 US tax law, real economic output, consumption, business investment, and payrolls grew faster than pre-law forecasts and nearly matched post-law projections, and an analysis of administrative corporate tax return data estimates firms with the average tax change raised domestic investment about 20% relative to firms with no change.
  • The Budget and Economic Outlook: 2025 to 2035 — The Congressional Budget Office projects the federal deficit will grow from $1.9tn (6.2% of GDP) to $2.7tn by 2035, well above the 3.8% of GDP average of the past 50 years. Outlays are projected to rise to 24.4% of GDP, driven by growth in Social Security and Medicare spending and rising net interest costs, while revenues reach 18.3%.
  • Interest Costs Would Explode with $5.5 Trillion Reconciliation Package — Analysts estimate that enacting $5.5tn of unoffset tax cuts and spending would raise federal interest costs by about $1.3tn over a decade, lifting annual net interest payments to $2tn, or 4.7% of GDP, by 2035 — above the prior record of 3.2% — and interest as a share of revenue to 27%.

AI Summary. Chinese researchers say they have modelled how the V-shaped surface wake of a submerged submarine generates a faint magnetic field as disturbed seawater ions interact with the Earth's geomagnetic field. They calculate that a Seawolf-class submarine at 24 knots and 30m depth produces a field of 10⁻¹² tesla, which they say is within the sensitivity range of existing airborne magnetometers.

Stephen Chen South China Morning Post
Date Posted:

Chinese researchers claim to have developed a method for tracking submarines via their magnetic fields. “The future lies in merging physics with AI. Magnetic wakes add another layer to the kill web – a system where even ghosts leave traces.” @SCMPNews

Can submarine surface wakes create magnetic signals detectable from aircraft?

Core argument: Researchers at Northwestern Polytechnical University in Xian calculate that a Seawolf-class submarine travelling at 24 knots (12.5 metres per second) and 30 metres depth generates a wake magnetic field of 10⁻¹² tesla, which they say falls within the sensitivity range of existing airborne magnetometers.

Researchers with Northwestern Polytechnical University (NPU) in Xian claim to have developed a groundbreaking method to detect even the quietest underwater vessels by harnessing the magnetic fields generated by their wakes – a discovery that could reshape naval warfare. Led by associate professor Wang Honglei, the team modelled the Kelvin wake, a V-shaped surface disturbance created by submarines as they slice through water. This wake, previously studied for radar-based imagery detection, generates a faint but detectable magnetic field when seawater ions – disturbed by the vessel’s motion – interact with the Earth’s geomagnetic field. For a Seawolf-class submarine travelling at 24 knots (12.5 metres per second) and 30 metres (98 feet) depth, the wake’s magnetic field reaches 10⁻¹² tesla – “well within the sensitivity range of existing airborne magnetometers”, according to Wang and his colleagues.

Takeaways by Macro Roundup® AI

  1. Researchers at Northwestern Polytechnical University in Xian calculate that a Seawolf-class submarine travelling at 24 knots (12.5 metres per second) and 30 metres depth generates a wake magnetic field of 10⁻¹² tesla, which they say falls within the sensitivity range of existing airborne magnetometers.
  2. The NPU team, led by associate professor Wang Honglei, modelled the V-shaped Kelvin wake created by a submarine's motion, which generates a faint magnetic field as disturbed seawater ions interact with the Earth's geomagnetic field.
  3. Wang and his colleagues claim a method to detect even the quietest underwater vessels, extending Kelvin wake analysis beyond its prior study for radar-based imagery detection.

Related Articles:

  • The Era of Total U.S. Submarine Dominance Over China Is Ending — China has launched a nuclear-powered attack submarine using pump-jet propulsion, the first appearance on a Chinese submarine of noise-reducing technology used on the newest American submarines, and satellite images show expanded submarine construction capacity. China has also built or nearly completed underwater sensor networks around its coast that improve its ability to detect enemy submarines.
  • US Nuclear Submarine Weak Spot In Bubble Trail: Chinese Scientists — Researchers at the Chinese Academy of Sciences calculate that the extremely low frequency magnetic signal generated by the bubble trail of an advanced submarine could exceed the sensitivity of advanced magnetic anomaly detectors by three to six orders of magnitude. The authors argue this offers a new approach to detecting and tracking submarines from long distances.
  • China’s Newest Nuclear Submarine Sank, Setting Back Its Military Modernization — U.S. officials say China's first Zhou-class nuclear-powered attack submarine, which features a distinctive X-shaped stern designed for greater maneuverability, sank at a shipyard near Wuhan, a setback for a priority weapons program that Chinese authorities sought to conceal.