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
  • “Unintended Consequences offers deep and well-argued analyses on almost every issue.” - The New York Times
  • “…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
  • “Unintended Consequences represents the most cogent and persuasive analysis of the Financial Crisis to date.” - Andrei Shleifer, 1999 John Bates Clark Medal Winner
  • “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 comprehensive explanation of the modern economy.” - Julian Robertson, Founder, Tiger Management
  • “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
  • “…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
  • “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
  • “…challenges misconceptions that distort our economic debates.” - Arthur Brooks, President of the American Enterprise Institute
Upside of Inequality Oxford Unintended Consequences

Macro Roundup


Monday, July 6, 2026

AI Summary. Analysis of US stock portfolios from 1971 to 2025 finds that "do-nothing" portfolios, which keep holding stocks even after they leave the S&P 500 index, on average match index returns, and equal-weighted versions outperformed market-cap-weighted ones before trading costs (14.02% vs 12.69% average 1-year returns). Portfolios built from the largest constituents underperformed those built from all constituents over the period.

Hendrik Bessembinder Arizona State University
Date Posted:

Over 1971–2025, the equal-weighted S&P 500, and similar portfolios with a larger small-cap tilt, returned 1 to 2pp more per year than the value-weighted S&P. “The outperformance of the largest stocks over the past dozen or so years is…atypical.”

Do buy-and-hold stock portfolios match index returns?

Core argument: Equal-weighted "do-nothing" portfolios returned an arithmetic mean of 14.02% over 1971–2025 versus 12.69% for value-weighted portfolios, with geometric means of 12.59% and 11.30% respectively, before trading costs.

I assess the impact during the 1971 to 2025 period of selecting portfolios based on equal-weighting versus value-weighting [i.e.weighting by market cap]. The data in Table 1 indicates that equal-weighted “do-nothing” portfolios outperform the value-weighted do-nothing portfolios overall. The arithmetic mean of 1-year-equal weighted portfolio returns is 14.02% while that of the value-weighted 1-year portfolio returns is 12.69%. The corresponding geometric mean returns are 12.59% for the equal-weighted portfolio compared to 11.30% for the value-weighted portfolio. “Do Nothing” portfolios, which continue to hold stocks even if they delist from the S&P 500 index, on average generate returns that match the index itself. On the other hand, equal weighted “Do Nothing” portfolios on-average outperformed value-weighted, at least before taking trading costs into account. Randomly selected narrow portfolios generate average returns similar to portfolios constructed from all constituent stocks, but underperform more than half of the time over longer holding periods. Strikingly, “Do Nothing” portfolios constructed from the largest S&P constituent firms have strongly underperformed portfolios constructed from all constituents over the past 55 years, especially when the investment is entirely in the largest stock. The outperformance of the largest stocks over the past dozen or so years is, in fact, atypical.

Takeaways by Macro Roundup® AI

  1. Equal-weighted "do-nothing" portfolios returned an arithmetic mean of 14.02% over 1971–2025 versus 12.69% for value-weighted portfolios, with geometric means of 12.59% and 11.30% respectively, before trading costs.
  2. "Do-nothing" portfolios that keep holding stocks after they exit the S&P 500 generate average returns matching the index itself.
  3. "Do-nothing" portfolios built from the largest S&P constituents strongly underperformed portfolios built from all constituents over the past 55 years, especially when the investment is entirely in the single largest stock.

Related Articles:

  • One Hundred Years in the U.S. Stock Markets — Over a century, U.S. publicly traded stocks generated roughly $91tn in shareholder wealth enhancement, with returns strongly positively skewed and dominated by a relatively small number of high-performing firms. Because most possible outcomes fall below the mean in a skewed distribution, the authors argue financial planning should focus on median outcomes or the full distribution.
  • Opportunities and Expectations: The Present Value of Growth Opportunities in Valuation — A stock's value can be split into steady-state value from current operations and the "present value of growth opportunities" from future value-creating investments, which has averaged 35% of the S&P 500's price historically. Lower shares of price attributed to growth opportunities have been associated with higher subsequent returns, most clearly at the extremes.
  • Competitive Advantage Period: The Neglected Value Driver — Data indicate persistence in returns on invested capital has increased since 2000, with large companies the main beneficiaries; the information technology sector's aggregate return was 19% versus a 10.8% median from 1970 to 2024. Of about 13,800 U.S. public firms listing from 1976 to 2019, 18% survived.

AI Summary. Productivity growth from 2000-2019 was higher in the US than in western Europe, but almost half the gap disappears when information and communications technology is excluded, even though tech was 9.2% of US output versus 5.4% in the EU. The author estimates this difference in industrial mix accounts for roughly 40% of the overall productivity growth gap.

Paul Krugman Krugman Wonks Out
Date Posted:
Is Database:
Database
Is Important:
Important

The superior productivity growth of the US relative to Europe from 2000–2019 is highly concentrated in the tech sector. Krugman computes the productivity growth differential excluding ICT and finds that it shrinks from .65 to .38 per year.

How much of the US-Europe productivity gap reflects tech sector size?

Core argument: Productivity growth over 2000–2019 was higher in the US than in the EU11, essentially western Europe, but almost half of that gap disappears when the ICT sector is excluded.

The [Draghi] report states that, apart from the ICT [information, communication and technology] sectors in the EU and the US, productivity performance across the rest of the two economies has been similar. Chart 5 shows what I find for productivity growth rates from 2000-2019 in the US and the EU11, basically western Europe. Overall productivity growth was higher in the US than in Europe, but almost half the difference goes away if one excludes ICT, even though the tech sector was only 9.2% of US GDP and 5.4% of EU GDP. It’s worth noting that, within the US, productivity rose much faster in tech than in the rest of the economy. When trying to model the US-EU paradox, I focused on the difference in the share of GDP accounted for by the tech sector. As noted above, the productivity gap outside tech is relatively small. How much of the US-EU paradox can this story, which emphasizes differences in industrial mix, explain? Industrial mix effect = (Share of tech in US GDP – share of tech in EU GDP) * (productivity growth rate in tech – productivity growth rate in non-tech). If I plug in the US-EU difference in ICT share — 3.8 percentage points — and the difference within the US between productivity growth in ICT and in the rest of the economy — 6.8 percentage points — I get 0.26 percentage points, or approximately 40% of the reported overall US-EU difference in productivity growth.

Takeaways by Macro Roundup® AI

  1. Productivity growth over 2000–2019 was higher in the US than in the EU11, essentially western Europe, but almost half of that gap disappears when the ICT sector is excluded.
  2. The tech sector accounted for 9.2% of US GDP versus 5.4% of EU GDP, a 3.8-percentage-point difference, while US productivity growth in ICT exceeded that in the rest of the US economy by 6.8 percentage points.
  3. Multiplying the 3.8-point ICT share gap by the 6.8-point within-US productivity growth gap yields an industrial mix effect of 0.26 percentage points, roughly 40% of the reported overall US-EU productivity growth difference.

Related Articles:

  • 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.
  • European Stagnation Is Real — Comparing the United States, France, and Germany at current purchasing power parity prices shows the European economies roughly constant relative to America since 2000, but current-price measures miss output gains in sectors where prices fall, such as software. The author argues constant-price comparisons, the standard approach for growth over time, reveal greater European stagnation.
  • Is Europe in Economic Decline? — French gross domestic product per person, adjusted for purchasing power, was 74% of the US level in 2001 and 73% in 2024, while Germany closed some of the gap, indicating European living standards have not fallen behind. Higher measured US growth is attributed to rapid productivity growth in information technology, a sector the US dominates.

AI Summary. Emerging-market central bank reserve diversification into gold, following the freezing of Russia's reserves, anchors a forecast of $4,900 per troy ounce by end-2026, with risks skewed to the upside. Near-term headwinds include reduced demand for currency-debasement hedges and rate-sensitive exchange-traded fund demand if the Federal Reserve turns hawkish.

Samantha Dart, Daan Struyven, Lina Thomas, Yulia Zhestkova Grigsby, et al. Goldman Sachs
Date Posted:

Gold is up 123% since 2022. GS argues it still has room to run, due to EM central bank buying and private diversification driven by worries about Western fiscal sustainability. They have a price target of $4,900/toz by the end of 2026.

Does central bank reserve diversification support further gold price gains?

Core argument: Gold has risen 123% since 2022, and the authors forecast $4,900/toz by end-2026, anchored structurally by emerging-market central bank reserve diversification following the 2022 freezing of Russia's reserves.

Despite gold’s strong rally since 2022 (up 123%), we continue to see further upside, driven by both structural and eventually cyclical factors. Structurally, EM central bank diversification--following the 2022 freezing of Russia’s reserves--remains the anchor of our $4,900/toz end 2026 forecast. While the pace of central bank gold purchases has moderated to ~50 tonnes/month on a 3-month (seasonally adjusted) and 12-month moving average basis, we view the ongoing diversification trend as structural. Cyclically, however, gold faces near term headwinds from the potential hit to demand for macro policy hedges (as a hawkish Fed helps fade the debasement theme) and as markets are pricing in Fed hikes this year amid inflation concerns, weighing on rate-sensitive ETF demand. We expect these headwinds to at least partly reverse over time. Over the medium term, risks to our gold price forecast remain skewed to the upside on the net. Gold's share in private portfolios remains small, and the Iran episode -- together with broader geopolitical developments -- may eventually accelerate private diversification into gold, including by weighing on perceptions of Western fiscal sustainability.

Takeaways by Macro Roundup® AI

  1. Gold has risen 123% since 2022, and the authors forecast $4,900/toz by end-2026, anchored structurally by emerging-market central bank reserve diversification following the 2022 freezing of Russia's reserves.
  2. Central bank gold purchases have moderated to roughly 50 tonnes per month on both 3-month seasonally adjusted and 12-month moving average bases, a diversification trend the authors still view as structural.
  3. Gold faces near-term headwinds from a hawkish Federal Reserve fading the debasement theme and from markets pricing Fed hikes amid inflation concerns, weighing on rate-sensitive ETF demand.

Related Articles:

  • Gold Heads For Worst Quarter In More Than A Decade As Retail Frenzy Fades — Gold fell below $4,000 a troy ounce, down nearly 14% over three months, its worst quarterly performance in more than a decade after reaching a record high near $5,595. Analysts attribute the decline to expectations of US interest rate rises, outflows from gold-backed funds, traders shifting into AI stocks, and Chinese bank restrictions on retail bullion trading.
  • The Debasement Trade Is Unraveling and Kevin Warsh Is One Big Reason — The "debasement trade" — favoring gold and Bitcoin over a dollar seen as vulnerable to inflationary, fiscal and monetary excess — has been a defining market narrative, but a new Federal Reserve chairman's emphasis on price stability as the overriding priority lifted the dollar and long-dated Treasuries at the expense of gold and crypto.
  • Global Debt Report 2026 — Gross central government borrowing across OECD countries rose 7% to a record $17tn, about 23% of GDP, and is projected to reach roughly $18tn, or 24% of GDP. The United States and Japan account for close to 80% of the area's refinancing requirements, with the US share rising from 57% to 70%.

AI Summary. Mitsubishi Heavy Industries argues that car factories, built to produce large volumes of identical units, are ill-suited to military drones, whose specifications change constantly, while its own expertise lies in low-volume, high-variation production. The company says data security concerns will push Japan's defence ministry to source some drones domestically without international partners.

Harry Dempsey Financial Times
Date Posted:

Mitsubishi Heavy’s CEO argues that pivoting automotive manufacturing capacity to drones risks producing large numbers of quickly obsolete drones. “Products in this field change their specifications constantly depending on the situation.”

Can car factories be converted to build military drones?

Core argument: Mitsubishi Heavy Industries developed an interceptor drone prototype capable of destroying enemy unmanned aerial vehicles in three months, drawing on its expertise in satellites, command and control systems, and submarines.

Eisaku Ito, chief executive of Mitsubishi Heavy Industries, believes the firm's drone business would grow to a “considerable scale.” He argues the firm could become Japan’s premier drone supplier because it specialised in producing low volumes with high levels of variation. "Products in this field change their specifications constantly depending on the situation. Automobile factories, by contrast, are designed to manufacture tens of thousands or millions of the same units." The defence ministry would want Japan to develop and make some drones without international partners because of data security concerns, Ito said, giving MHI an advantage. “We’re the only company in Japan that can handle this comprehensively,” he said, referring to its ability to produce military kit covering sea, land, and air. In just three months, MHI recently developed an interceptor drone prototype that can take out enemy unmanned aerial vehicles, drawing on its expertise in satellites, command and control systems, and submarines.

Takeaways by Macro Roundup® AI

  1. Mitsubishi Heavy Industries developed an interceptor drone prototype capable of destroying enemy unmanned aerial vehicles in three months, drawing on its expertise in satellites, command and control systems, and submarines.
  2. Mitsubishi Heavy Industries chief executive Eisaku Ito argues that automobile factories are ill-suited to military drone production because drone specifications change constantly with the situation, while car plants are designed to make tens of thousands or millions of identical units.
  3. Eisaku Ito says Japan's defence ministry would want some drones developed and made without international partners because of data security concerns, giving Mitsubishi Heavy Industries an advantage as, in his words, the only Japanese company able to handle sea, land and air military kit comprehensively.

Related Articles:

  • Can Drones Be Stockpiled? Europe Wrestles With Weapons Dilemma — European defense officials argue that small, cheap drones have proven as effective as heavy armour in Ukraine, but stockpiling millions of them risks holding weapons made obsolete within months by new software, radio frequencies, and autonomous navigation. Drones that dominate one week can be jammed or outclassed the next.
  • Pentagon Approaches Automakers, Manufacturers to Boost Weapons Production — The Pentagon has approached U.S. automakers and other manufacturers about using their workers and factory capacity to boost production of munitions and other equipment as wars deplete stocks, asking whether they could rapidly shift to defense work. Most large non-defense manufacturers already hold Pentagon contracts, but these are typically limited in scope and value.
  • Why Every Country Needs To Master The Electric Tech Stack — Drones are the foundation of hard power in the early 21st century, so a country unable to produce them domestically is at a disadvantage. As electric motors overtake combustion engines as the core technology that makes machines move, the same supply chains and production processes used for electronics can also produce cars, drones, robots, and appliances.

AI Summary. Chinese policymakers are discussing reducing incentives for researchers to publish in foreign journals, including cutting the weight given to leading international journals in promotion and tenure decisions. China-based authors accounted for almost a third of papers in a database of leading international scientific journals, up from 5% two decades earlier.

Sun Yu, Michael Peel and Eleanor Olcott Financial Times
Date Posted:

China is seeking to restrict the publication of domestic scientific research abroad. The government may push Chinese universities and research institutions to “reduce or remove the weight given to publishing in leading international journals.”

Is China moving to discourage researchers from publishing in foreign journals?

Core argument: Papers by China-based authors made up almost a third of the global total in the SCI database of leading international scientific journals in 2024, up from 5% two decades earlier.

Chinese policymakers are discussing ways to reduce the incentives for scientific researchers to submit their findings to foreign journals. The government could push Chinese universities and research institutions to reduce or remove the weight given to publishing in leading international journals in deciding academic promotion and tenure, the people said. Papers by China-based authors accounted for almost a third of the global total in the SCI — a database of leading international scientific journals — in 2024, up from 5% two decades earlier.

Takeaways by Macro Roundup® AI

  1. Papers by China-based authors made up almost a third of the global total in the SCI database of leading international scientific journals in 2024, up from 5% two decades earlier.
  2. Chinese policymakers are discussing ways to reduce incentives for researchers to submit findings to foreign journals.
  3. One option under discussion is pushing Chinese universities and research institutions to reduce or remove the weight given to publication in leading international journals in academic promotion and tenure decisions.

Related Articles:

  • The Growing Self-Reliance of Chinese Innovation — Linking patents to scientific papers through shared phrases, the authors find China's innovation has grown markedly more self-reliant across nearly all technology areas the U.S. treats as strategically important. In an aggregate grouping of these areas, the share of patent phrases traced to Chinese-produced articles rose from 2% to 24%, while the U.S.-produced share fell from 31% to 18%.
  • China and the Future Of Science — China graduates about 5x as many medical and biomedical students as the US each year, 7x as many engineers, and nearly double the number of science and engineering PhDs, while one-sixth to one-fifth of US science and engineering graduates are themselves Chinese. One study found returning Chinese scientists lead-author 2.5x more papers than peers who stay in the US.
  • How China Is Outperforming the United States in Critical Technologies — The Australian Strategic Policy Institute's Critical Technology Tracker finds China leads in 57 of 64 critical technology categories while the United States leads in 7, and that China surpasses the United States in top-tier scientific publications across all 8 technology domains. In energy and environment China accounts for 46% of top-tier publications versus 10% for the United States.

Tuesday, July 7, 2026

AI Summary. A 1pp lower birth rate 20 years earlier is associated with 22 log points higher output per worker 20 years later and 29 log points higher after 40 years, about 0.73pp faster annual growth. The authors find lower birth rates are also associated with more labor-saving patents, larger high-tech shares, and higher productivity growth.

Daron Acemoglu, David Autor, Keelan Beirne and Andrew Scott National Bureau of Economic Research
Date Posted:
Is Database:
Database
Is Important:
Important

Cross-country evidence from 1950 to 2020 shows that a 1pp lower birth rate is associated with 22 log points (~25%) higher GDP per worker 40 years later and 29 log points (~44%) after 60 years, likely driven by labor-saving technical change.

Do lower birth rates raise output per worker decades later?

Core argument: A 1pp lower birth rate 20 years earlier is associated with 22 log points higher GDP per worker at t+20 and 29 log points by t+40, roughly 0.73 percentage points faster annual growth over the 40-year interval.

Figure 3 traces out how GDP per worker evolves following a 1pp increase in birth rates at t−20. Since cohorts born at t−20 (e.g., 1970) do not enter the working-age population until t = 0 (e.g., 1990), we should observe no response before t = 0; [indeed] the estimated coefficients are ~0 and statistically insignificant before t = 0. Beginning at t = 0, however, higher birth rates are associated with lower GDP per worker, and the negative effects grow progressively larger over time. For example, a 1pp lower birth rate at t−20 is associated with 22 log points higher GDP per worker at t+20 and 29 log points higher by t+40. The latter corresponds to roughly 0.73 percentage points faster annual growth over the 40-year interval. We find that lower birth rates are associated with more labor-saving patents and a growing share of high-tech industries across countries and US commuting zones [Figure 11]. They are also predictive of higher TFP growth across countries and US industries and increased patenting in ICT and broader automation technology classes. Using cross-country variation in WWII-era military and civilian deaths, we present suggestive evidence that the positive growth effects of falling birth rates are primarily driven by the scarcity of younger workers rather than by reductions in population per se [Figure 14]. [Editor’s note: the paper does not grapple with the fiscal consequences of an aging population.]

Takeaways by Macro Roundup® AI

  1. A 1pp lower birth rate 20 years earlier is associated with 22 log points higher GDP per worker at t+20 and 29 log points by t+40, roughly 0.73 percentage points faster annual growth over the 40-year interval.
  2. Estimated coefficients for GDP per worker are near zero and statistically insignificant before affected cohorts enter the working-age population at t = 0.
  3. Lower birth rates are associated with more labor-saving patents, a growing high-tech industry share across countries and US commuting zones, higher TFP growth, and more ICT and automation patenting, with WWII-era death variation offering suggestive evidence that scarcity of younger workers, rather than smaller population, drives the growth effects.

Related Articles:

  • Can A Depopulating America Still Flourish? — US labor force participation among women aged 25-54 has reached all-time highs, exceeding its previous 1999 peak, yet remains over 4 points below Europe's level and nearly 6 points below Japan's. About 3.5m prime-age American women are neither working nor looking for work, without children at home or husbands, compared with roughly 7m such men.
  • Capital Is Making a Comeback — Tangible investment in property, factories and equipment fell from 12.5% to 8.5% of private gross domestic product, while intangible investment in intellectual property, software and process knowledge rose from about 11.5% to 16.75%. The author argues demand for tangible capital investment will now have to rise.
  • 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.

AI Summary. Aging populations create structural incentives to automate, and the argument is that demographic change could become one of the strongest long-run drivers of artificial intelligence deployment. Technology-producing economies may benefit twice, offsetting their own labor shortages while exporting automation, robotics, and AI to other aging societies.

Paul Kedrosky Applied Complexity
Date Posted:

Kedrosky argues that labor scarcity may become a “permanent innovation driver” as capital increasingly substitutes for labor, driving automation. “AI adoption may be driven as much by labor scarcity as by breakthroughs in machine learning.”

Are aging populations becoming a leading driver of AI adoption?

Core argument: Kedrosky argues that aging societies create structural incentives to automate, which could make demographic change one of the strongest long-run drivers of AI deployment.

[Commenting on Autor's new research, Kedrosky argues] AI adoption will be driven by demographic change. Aging societies create structural incentives to automate. Demographics could become one of the strongest long-run drivers of AI deployment. The economic winners of aging will be technology-producing economies. Countries that develop automation, robotics, and AI may benefit twice. First, by offsetting their own labor shortages. Second, by exporting those technologies to every other aging society. Demographics could reinforce, rather than weaken, technological concentration. Measures of national economies may shift from scale to productivity. If aggregate GDP becomes less useful while output per worker continues to rise, economies may increasingly be judged by productivity, innovation, and living standards rather than by the size of their labor force or total output. Labor scarcity becomes a permanent innovation driver. Tight labor markets cease to be a short-term inflationary problem and instead direct capital toward automation. Persistent worker shortages could accelerate AI adoption across sectors that previously resisted technological change, like healthcare, compressing decades of automation into a few years.

Takeaways by Macro Roundup® AI

  1. Kedrosky argues that aging societies create structural incentives to automate, which could make demographic change one of the strongest long-run drivers of AI deployment.
  2. Kedrosky contends that technology-producing economies may benefit twice from aging, offsetting their own labor shortages and exporting automation, robotics, and AI to other aging societies, potentially reinforcing technological concentration.
  3. Kedrosky argues that persistent worker shortages direct capital toward automation rather than acting as a short-term inflationary problem, and could accelerate AI adoption in sectors that previously resisted it, such as healthcare.

Related Articles:

  • Baby Busts and Growth Booms: Demographic Change and the Macroeconomy — A 1pp lower birth rate 20 years earlier is associated with 22 log points higher output per worker 20 years later and 29 log points higher after 40 years, about 0.73pp faster annual growth. The authors find lower birth rates are also associated with more labor-saving patents, larger high-tech shares, and higher productivity growth.
  • Capital Is Making a Comeback — Tangible investment in property, factories and equipment fell from 12.5% to 8.5% of private gross domestic product, while intangible investment in intellectual property, software and process knowledge rose from about 11.5% to 16.75%. The author argues demand for tangible capital investment will now have to rise.
  • The Demographic Outlook: 2026 to 2056 — The Congressional Budget Office projects the US population rises from 349m to 364m between 2026 and 2056, with annual growth slowing from 0.3% over the next decade to 0.1% thereafter before reaching zero and then shrinking. Declining fertility is projected to push deaths above births starting in 2030, increasingly offsetting net immigration.

AI Summary. Europe consumes AI products like ChatGPT without capturing the capital gains, payroll taxes on high-paid engineers, or agglomeration effects, because the tax base accrues where high-value production occurs rather than where products are consumed. The author argues Europe's welfare commitments were implicitly backed by expected future growth that is no longer anticipated.

Hanno Lustig The Two Cents
Date Posted:
Is Database:
Database

Lustig argues that Krugman is wrong to dismiss concerns about slower European growth relative to the US based on welfare measures: “Europe’s social model is not financed out of consumer surplus; it is financed out of taxable income and taxable profits.”

Does consuming AI without producing it erode Europe's tax base?

Core argument: Europe consumes ChatGPT but collects none of the capital gains, payroll taxes on million-dollar AI engineers, or agglomeration effects, because the tax base migrates to where high-value production occurs rather than where products are consumed.

Europe consumes ChatGPT; it does not collect the capital gains, the payroll taxes on million-dollar AI engineers, or the agglomeration effects. The worst possible combination for a welfare state: contented consumers committed to the status quo, eroding public finances. The tax base migrates to where the high-value production takes place, not where the products are ultimately consumed. All this insurance that European states are committed to providing was implicitly backed by the promise of future growth. And that growth is no longer expected to materialize, but the promises remain. The problem is not simply that Europeans chose a different point on the efficiency-equity frontier. Their choices have shifted the frontier itself inward over time, leaving future generations of Europeans with a trade-off that is much worse.

Takeaways by Macro Roundup® AI

  1. Europe consumes ChatGPT but collects none of the capital gains, payroll taxes on million-dollar AI engineers, or agglomeration effects, because the tax base migrates to where high-value production occurs rather than where products are consumed.
  2. Europe's welfare-state insurance was implicitly backed by a promise of future growth that is no longer expected to materialize, while the promises themselves remain in place.
  3. European choices have shifted the efficiency-equity frontier itself inward over time, leaving future generations with a worse trade-off rather than merely a different point on the same frontier.

Related Articles:

  • European vs. U.S. Economic Performance: An Update — Productivity growth from 2000-2019 was higher in the US than in western Europe, but almost half the gap disappears when information and communications technology is excluded, even though tech was 9.2% of US output versus 5.4% in the EU. The author estimates this difference in industrial mix accounts for roughly 40% of the overall productivity growth gap.
  • 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.
  • European Stagnation Is Real — Comparing the United States, France, and Germany at current purchasing power parity prices shows the European economies roughly constant relative to America since 2000, but current-price measures miss output gains in sectors where prices fall, such as software. The author argues constant-price comparisons, the standard approach for growth over time, reveal greater European stagnation.

AI Summary. An index blending artificial intelligence token prices and usage is ambiguous when it softens, since token prices have fallen more than 90% since 2023 while total spending has roughly doubled over the past year. Allianz Research measures a nearly 46% growth gap between artificial intelligence investment and sales, wider than the 32% divergence during the 2001 telecom bust.

Jan-Patrick Barnert and Michael Msika Bloomberg
Date Posted:
Is Database:
Database

The Silicon Data LLM Token Expenditure Index, which tracks AI token prices, is down ~20% from its recent peak in May. The metric is a rough “proxy for marginal willingness to pay.”

Why does a softening AI token index send ambiguous signals?

Core argument: Token prices have fallen more than 90% since 2023 while total token spend has roughly doubled over the past year, supporting a benign read in which cheaper tokens expand the market and an index pause reflects digestion.

A softer index doesn’t mean AI is getting cheaper. The gauge blends prices and usage, meaning a dip can imply very different scenarios: either list prices are falling, or demand is shifting toward cheaper models. It could also point to a genuine softening in what buyers are prepared to shoulder. Each of these possibilities carries different implications. Let’s start with a benign read: While token prices have collapsed more than 90% since 2023, total spend has roughly doubled since last year. Cheaper tokens have expanded the market. This means that an index pause is simply digestion, while demand is real and capex is money well spent. The bull case for Nvidia rests here. Now for the interpretation that’s keeping people up at night: Bears warn that sustained weakness in the index could end the trade that saw nearly the entire AI cohort rally hard this cycle. It’s token spending that justifies the next capex order, and the bill is already looking stretched. Allianz Research said there’s nearly a 46% growth gap between AI investment and sales. That’s worse than the 32% divergence measured during the 2001 telecom bust.

Takeaways by Macro Roundup® AI

  1. Token prices have fallen more than 90% since 2023 while total token spend has roughly doubled over the past year, supporting a benign read in which cheaper tokens expand the market and an index pause reflects digestion.
  2. Allianz Research put the growth gap between AI investment and sales at nearly 46%, wider than the 32% divergence measured during the 2001 telecom bust.
  3. The index blends prices and usage, so a dip can imply falling list prices, demand shifting toward cheaper models, or a genuine softening in what buyers are prepared to shoulder.

Related Articles:

  • AI: The ROI Runway Could Be Long Outside the Tech Sector — AI company valuations rest on the assumption that profit margins across non-tech firms will eventually rise, but outside a handful of sectors adoption is slow, creating a divergence between front-loaded valuations and slower cash flows. If the per-use costs of AI models converge toward zero, revenue may be insufficient for all major cloud providers.
  • 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.
  • Chart of the Day: Small Models are Closing the Gap to Frontier AI — Achieving 60% on MMLU, a test of academic reasoning, once required a 540bn-parameter model but was reached by a 3.8bn-parameter model two years later, a 142x reduction in parameters. Models small enough to run on phones or sensors shift inference from data centers to devices, weakening the case for paying cloud providers per token.

AI Summary. Tech-related investment rose 30% from year-ago levels while all other capital spending fell, a divergence the author argues reflects the AI boom bidding away scarce physical inputs such as electricians, grid capacity, metals, and construction labor. With hyperscalers' free cash flow now exhausted, over 80% of projected capital spending in coming years will require external financing.

Jason Thomas Carlyle
Date Posted:
Is Database:
Database

Tech-related net fixed investment had risen 30% y/y by the end of H1, while non-tech investment fell, suggesting the datacenter buildout is crowding out non-tech related fixed investment.

Is the AI investment boom crowding out other capital spending?

Core argument: Tech-related investment rose 30% from year-ago levels while all other capex fell, a record divergence implying the AI boom is bidding away electricians, grid capacity, transformers, metals, engineering talent, and construction labor in quantities that stress rival projects' economics.

Overall tech-related investment increased 30% from year-ago levels, while all other capex fell. This record divergence implies that the AI boom is bidding away physical inputs – electricians, grid capacity and transformers, metals and materials, engineering talent, and construction labor – in quantities that stress the economics of rival projects elsewhere. Perhaps the best illustration of binding real resource constraints was observed inside the AI data center complex itself. In the first half of the year, the average price of semiconductors, labor for assembly and integration, and related materials exploded, as physical supply constraints interacted with insatiable demand. As more economic rent accrued to the compute layer, the stock prices of the semiconductor manufacturers upstream from this physical chokepoint outperformed those of their top customers by 56%. When AI spending gets financed out of internally generated cash flow, “crowding out” isn’t a pressing issue for investors. Credit spreads have not widened materially, financing conditions remain accommodative, and the cost of capital for non-software borrowers doesn’t suggest displacement. But the hyperscalers’ free cash flow has now been exhausted. Over 80% of the projected capex over the next few years will have to be externally financed, including fresh equity issuance.

Takeaways by Macro Roundup® AI

  1. Tech-related investment rose 30% from year-ago levels while all other capex fell, a record divergence implying the AI boom is bidding away electricians, grid capacity, transformers, metals, engineering talent, and construction labor in quantities that stress rival projects' economics.
  2. Semiconductor manufacturers' stock prices outperformed those of their top customers by 56% as more economic rent accrued to the compute layer, with average prices for chips, assembly and integration labor, and related materials rising sharply in the first half of the year.
  3. Hyperscalers' free cash flow has been exhausted, and over 80% of projected capex over the next few years will have to be externally financed, including fresh equity issuance, though credit spreads have not widened materially.

Related Articles:

  • The Macro Implications of the AI Capex Boom — Artificial intelligence capital spending is estimated to add roughly 1.4 percentage points to US growth in one year and about 1.5 points the next, comparable to business investment's contribution during the tech bubble. A material part of that estimate reflects profits at Nvidia and a few other chipmakers that are unlikely to be recycled back into the economy.
  • The Other US Capex Question — Business investment outside artificial intelligence has been weak, contributing just 1.54% to total capital expenditure growth. The author argues AI investment is not crowding out other projects, since the spread between return on capital and cost of capital has stayed positive and US nonfinancial corporations' capital spending is below their gross savings.
  • America’s Data Center Build-Out Is Falling Way Behind Schedule — A JPMorgan analysis found that more than 60% of US data-center capacity planned for completion in 2027 has not yet begun construction, with another 7% delayed. Analysts and power-industry experts say Google's strategy of owning generation—including its $4.75bn purchase of wind and solar developer Intersect—and shifting computing loads to follow power supply could speed its grid connections.