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


Wednesday, June 24, 2026

AI Summary. 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.

Michael Cembalest J.P. Morgan
Date Posted:

Cembalest notes the “cost vs performance tradeoffs appear to provide ample incentive for open model adoption” as usage costs for frontier models rise. Enterprise tokens, and in turn models, will likely be commoditized.

Will rising token prices push enterprises toward cheaper open models?

Core argument: Cursor and Anthropic charge 6x more for tokens delivered 2.5x faster, indicating some enterprises will pay a premium for ultra-fast tokens.

The frontier labs are shifting to usage-based token pricing. It’s expected to become the norm, at least for enterprises, on a cost-plus basis that includes total costs of ownership. This will increase frontier lab token prices, a trend that is already underway. As a result, some enterprises may seek to avoid frontier lab token price increases by migrating to cheaper open models. Frontier models are here to stay since they’re needed for cybersecurity, distillation training, agents, and scientific discovery, all of which require maximum performance. There are enterprises even willing to pay up for ultra-fast tokens. For example, Cursor and Anthropic charge 6x more for tokens that are 2.5x faster. That said, many tokens consumed in the future may not come from frontier models but from smaller open models that are up to the tasks.

Takeaways by Macro Roundup® AI

  1. Cursor and Anthropic charge 6x more for tokens delivered 2.5x faster, indicating some enterprises will pay a premium for ultra-fast tokens.
  2. Frontier labs are shifting to usage-based token pricing on a cost-plus basis that includes total cost of ownership, which is expected to raise frontier token prices, a trend already underway.
  3. Frontier models remain necessary for cybersecurity, distillation training, agents, and scientific discovery, though many tokens consumed in the future may come from smaller open models that enterprises adopt to avoid frontier price increases.

Related Articles:

  • 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.
  • Commoditization, Orchestration, and the New AI Stack — The author argues that as incremental gains from larger AI training runs shrink, the return on additional computing power falls, weakening the case for the training spending that dominates data center demand. Survivors of model commoditization will be those owning the orchestration layer or the applications where outcomes are priced.
  • The Geopolitics of AI: Decoding the New Global Operating System — The performance gap between the top American and Chinese AI models narrowed from 9.3% to 1.7%, with Chinese firms relying on cheaper open-source models—DeepSeek prices usage 20-40x below OpenAI's ChatGPT. This price advantage could win adoption in developing countries and build dependency on Chinese AI ecosystems.

AI Summary. Artificial intelligence models' capability gains are shrinking in percentage terms, falling from about 15% relative to the baseline early on to roughly 10% for the same absolute increment, narrowing the gap between frontier and good-enough models. The author argues this points to a commoditizing market where price, not minor model differences, becomes the main differentiator, pressuring frontier firms' margins.

Paul Kedrosky Applied Complexity
Date Posted:

Citing the Epoch’s Capabilities Index, Kedrosky notes “the gap from the model frontier to the 90th percentile is under 6%. The gap from the #1 to the #2 model is roughly zero. The premium that the best models command [are] collapsing with declining variance.”

Do shrinking relative AI capability gains turn models into a commodity?

Core argument: AI model gains that once delivered about 15% improvement relative to the baseline now buy roughly 10%, so progress continues while its rate declines.

Even if the absolute capability gain per unit of time is constant, the percentage gain is not. Early in the data, models improved by about 15% relative to the baseline. At the end of the data, the same increment buys a 10% relative gain. Progress is real, but its rate is declining. The economic implications are clear. When the variance is wide, being the frontier model meant something. You could, after all, be almost 50% better than the average model. And you could, in theory, price like it. The distance between frontier and good-enough was wide enough that it looked like an early moat. This is what maturing commodity markets look like. Price will become the main differentiator, not minor model differences. This will have huge implications, from the continuing and inexorable rise of China in AI, as we are already seeing with DeepSeek's massive pricing advantage, to the resulting margin pressure on frontier companies, and the implications for said companies' post-IPO performance.

Takeaways by Macro Roundup® AI

  1. AI model gains that once delivered about 15% improvement relative to the baseline now buy roughly 10%, so progress continues while its rate declines.
  2. Wide variance across AI models once let a frontier model be nearly 50% better than the average and, in theory, price accordingly, a gap that looked like an early moat.
  3. The author argues AI is maturing into a commodity market where price, not minor model differences, becomes the main differentiator, citing DeepSeek's pricing advantage, margin pressure on frontier companies, and implications for their post-IPO performance.

Related Articles:

  • 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.
  • Four Takeaways from Microsoft’s Satya Nadella at WEF Today — Microsoft's chief executive says the price of AI model usage ("tokens") falls by roughly half every three months. Commentary argues that such rapid deflation pulls demand forward, encourages overbuilding, and commoditizes models, shifting competitive advantage to orchestration and context rather than the model layer.
  • Commoditization, Orchestration, and the New AI Stack — The author argues that as incremental gains from larger AI training runs shrink, the return on additional computing power falls, weakening the case for the training spending that dominates data center demand. Survivors of model commoditization will be those owning the orchestration layer or the applications where outcomes are priced.

AI Summary. The share of US national income paid to workers fell 1.6 percentage points below its pre-pandemic level, reaching a post-war low. The authors find this decline stems mainly from changes within industries rather than shifts in activity across sectors, following cyclical patterns seen in earlier recessions.

Richard Audoly, Miles Guerin, Srinidhi Narayanan, and Rachel Schuh Federal Reserve Bank of New York
Date Posted:
Is Database:
Database

Labor’s share of income is at a post-war low. The 1.6pp post-pandemic decline in the labor share has primarily been driven by within-industry changes as firms improved profits.

Why has the labor share of income fallen to postwar lows?

Core argument: After stabilizing in the 2010s, the labor share declined again during the post-COVID period, falling 1.6 percentage points below its pre-pandemic level and standing at an all-time low for the post-war period.

After stabilizing in the 2010s, the labor share declined again during the post-COVID period, ultimately falling 1.6 percentage points below its pre-pandemic level. The labor share now stands at an all-time low in the post-war period. Labor share’s trajectory post-COVID broadly follows the cyclical patterns observed in earlier recessions, with a decline during the recovery phase that mirrors historical dynamics. The decline in the labor share since COVID is driven primarily by within-industry changes rather than shifts in economic activity across sectors. Taken together, these results suggest that the post-COVID decline follows the same cyclical patterns as earlier recessions and is driven by the same within-industry forces, and they provide little evidence that it will evolve differently from past episodes.

Takeaways by Macro Roundup® AI

  1. After stabilizing in the 2010s, the labor share declined again during the post-COVID period, falling 1.6 percentage points below its pre-pandemic level and standing at an all-time low for the post-war period.
  2. The post-COVID decline in the labor share is driven primarily by within-industry changes rather than shifts in economic activity across sectors.
  3. The labor share's post-COVID trajectory broadly follows the cyclical patterns of earlier recessions, including a decline during the recovery phase, with little evidence that it will evolve differently from past episodes.

Related Articles:

  • The Record Divide Between Corporate Profits and Worker Pay — Labor's share of gross domestic income fell to 51%, the lowest since records began in 1947, while profits' share rose to 12.1%, the highest since 1950. Inflation-adjusted hourly wages are up 3% since the end of 2019, compared with a 50% rise in profits.
  • Why the Labor Share Keeps Falling: Taxes! — The measured corporate labor share fell from 63% in 1978 to 58% in 2017, but the authors estimate that tax-driven reclassification of wages as S corporation profit and the shift of high-labor-share firms into partnerships account for about 1.6 percentage points, roughly a third of the decline.
  • The Rise of Pass-Throughs and the Decline of the Labor Share — The authors estimate that reallocating business activity to its pre-1986 organizational form accounts for one-third of the decline in the corporate sector labor share, which fell from 62.9% to 57.9% between 1978 and 2017. US tax reform encouraged firms to organize as pass-through entities, letting income be characterized as profits rather than labor payments.

AI Summary. Nine Federal Reserve policymakers projected an interest-rate increase in 2026 and the median projection for core inflation in 2027 rose to 2.5%, a hawkish surprise that lifted market pricing for the end-2026 policy rate by 0.2 percentage points. The authors still expect the Federal Reserve to leave its policy rate unchanged this year.

David Mericle Goldman Sachs
Date Posted:

While market pricing for FFR hikes by the end of 2026 increased, Goldman’s “base case is still that the FOMC will leave the policy rate unchanged this year.”

Do hawkish central bank projections signal a rate hike this year?

Core argument: Nine FOMC participants projected a rate hike in 2026 at the June meeting, versus Goldman Sachs' expectation of three, and the median 2027 Q4/Q4 core PCE inflation projection rose to 2.5% against an expected 2.3%.

The June FOMC meeting delivered a hawkish surprise, with nine participants projecting a hike in 2026 (vs. our expectation of three) and the median projection for 2027 Q4/Q4 core PCE inflation rising to 2.5% (vs. our expectation of 2.3%). Market pricing for the fed funds rate at the end of 2026 rose by 20bp in response to the projections and Chairman Kevin Warsh’s press conference. The FOMC meeting raises the risk of interest rate hikes later this year, but our base case is still that the FOMC will leave the policy rate unchanged this year. A majority could support a hike, perhaps for a range of different reasons, if upcoming inflation prints are uncomfortable and job growth remains solid. But we suspect that most of the voters still lean toward an unchanged policy rate and that participants treated the news about a deal with Iran and the reopening of the Strait of Hormuz cautiously for now.

Takeaways by Macro Roundup® AI

  1. Nine FOMC participants projected a rate hike in 2026 at the June meeting, versus Goldman Sachs' expectation of three, and the median 2027 Q4/Q4 core PCE inflation projection rose to 2.5% against an expected 2.3%.
  2. Market pricing for the fed funds rate at the end of 2026 rose 20bp following the June projections and Chairman Kevin Warsh's press conference.
  3. Goldman Sachs' base case remains that the FOMC leaves the policy rate unchanged this year, though a majority could support a hike if upcoming inflation prints are uncomfortable and job growth remains solid.

Related Articles:

  • The Case for Higher U.S. Rates — Underlying U.S. inflation continues to run faster than the Federal Reserve's target and may be accelerating slightly, while growth, consumer spending, and a healthy job market keep wages rising about 1 percentage point faster than in the years just before the pandemic. On that evidence, the author argues rate increases, not cuts, are the appropriate policy.
  • Home Alone: Inflation And The New Fed Chair — Current readings on labor market tightness, manufacturing price pressures, supply chain tightness and the output gap resemble conditions that have historically preceded Federal Reserve rate increases rather than cuts, and futures markets are pricing in hikes. Averaging several monetary policy rules implies a policy rate of 4.00%–4.85% versus the current 3.50%–3.75%.
  • US Workers Lose Buying Power After War Drives Up Prices — Inflation has outstripped wage growth in key private-sector industries including healthcare, finance, and retail for two straight months, largely because of a surge in energy prices driven by the Iran war. Gasoline prices have fallen since an interim US-Iran peace deal but remain nearly $1 a gallon above pre-conflict levels.

AI Summary. 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.

Ye Xie, Anya Andrianova, and George Lei Bloomberg
Date Posted:

Despite the lack of fiscal consolidation, the “Debasement Trade” is in retreat; traders are now pricing in two rate hikes by the end of Q1 2027. A troy ounce of Gold has fallen ~19% since it’s peak, but remains up ~60% since January 2025.

Does an inflation-focused central bank leader strengthen the dollar over gold?

Core argument: Kevin Warsh, leading his first policy meeting as Fed chairman, said price stability was his overriding priority, sparking a further resurgence in the dollar and lifting inflation-sensitive long-dated Treasuries at the expense of gold and crypto.

The debasement trade — broadly defined as a strategy favoring assets such as gold and Bitcoin over currencies vulnerable to inflationary, fiscal and monetary excess like the dollar — had been one of the defining market narratives of the past two years. While Warsh got the nod from Trump by advocating for lower interest rates, it was his prior reputation as an inflation hawk that caused [investors] to hedge their bets when he was named. Warsh, leading his first policy meeting as Fed chairman, said price stability was his overriding priority [which] sparked a further resurgence in the dollar and gave a lift to inflation-sensitive long-dated Treasuries, at the expense of gold and crypto. Ray Dalio and Ken Griffin have argued that gold may ultimately prove safer than the dollar and warned that the trajectory of US debt risks a future fiscal crisis.

Takeaways by Macro Roundup® AI

  1. Kevin Warsh, leading his first policy meeting as Fed chairman, said price stability was his overriding priority, sparking a further resurgence in the dollar and lifting inflation-sensitive long-dated Treasuries at the expense of gold and crypto.
  2. The debasement trade — favoring assets such as gold and Bitcoin over currencies vulnerable to inflationary, fiscal and monetary excess like the dollar — had been one of the defining market narratives of the past two years.
  3. Ray Dalio and Ken Griffin have argued that gold may ultimately prove safer than the dollar and warned that the trajectory of US debt risks a future fiscal crisis.

Related Articles:

  • 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%.
  • Currency Moves Can Drive Rebalancing — Trade responds to exchange-rate changes with a lag of roughly 8 to 12 quarters, and after the dollar fell from its mid-1980s peak the US trade deficit dropped from about 3% to about 1% of GDP while Japan's surplus fell from 4% to 2%. The author argues currency appreciation, not fiscal consolidation, drove that rebalancing.
  • The Dollar has Further to Fall — The author argues the dollar has considerably further to fall, noting Federal Reserve data show its real value nearly two standard deviations above its average since 1973, a level matched only twice before, each followed by depreciations of 25-30%. Financing the $1.1tn US current account deficit requires matching net capital inflows.

AI Summary. Research from PwC finds the five largest US defence contractors had a combined backlog of undelivered orders of $1.36tn, up 24% from the prior year, as Lockheed Martin, Northrop Grumman, Raytheon and L3Harris invest billions of dollars expanding manufacturing. The US government has signed framework agreements with four companies to produce more than 10,000 low-cost cruise missiles over three years.

Christian Davies Financial Times
Date Posted:

PwC research estimates that major defense contractors have a $1.36T backlog of orders through 2025, an increase of 25% y/y.

Are US defense contractors expanding fast enough to clear record order backlogs?

Core argument: PwC research puts the combined backlog of undelivered orders at the five largest US defence primes at $1.36tn through 2025, up 24% from the prior year.

The backlog in US weapons manufacturing is expected to dominate a White House meeting on Wednesday between Donald Trump and the top American defence companies. Lockheed Martin broke ground last month on a new facility in Alabama, while fellow primary contractors Northrop Grumman, Raytheon and L3Harris are also investing billions of dollars in expanding their manufacturing footprints. According to PwC research the five largest so-called US defence primes had a combined backlog of undelivered orders of $1.36tn through 2025, up 24% on the year before. The government announced framework agreements last month with four companies — Anduril, CoAspire, Leidos and Zone 5 — to produce more than 10,000 low-cost cruise missiles over three years from 2027.

Takeaways by Macro Roundup® AI

  1. PwC research puts the combined backlog of undelivered orders at the five largest US defence primes at $1.36tn through 2025, up 24% from the prior year.
  2. Lockheed Martin broke ground on a new Alabama facility, while Northrop Grumman, Raytheon and L3Harris are investing billions of dollars in expanding their manufacturing footprints.
  3. The US government announced framework agreements with Anduril, CoAspire, Leidos and Zone 5 to produce more than 10,000 low-cost cruise missiles over three years from 2027.

Related Articles:

  • Trump Invokes Cold War-Era Law to Boost Munitions Production — The U.S. has invoked the Defense Production Act to expand munitions output amid doubts that American weapons makers can meet increased demand. U.S. officials say more than 1,000 long-range Tomahawk missiles and 1,500–2,000 air-defense interceptors have been fired in the war with Iran, and replacing those stockpiles could take up to six years.
  • China Restricts Trading With Some US Rare Earth Companies — China has added 10 American companies to its "entities list," restricting trading with firms central to US efforts to build a rare earth supply chain rivaling Beijing's, including USA Rare Earth, MP Materials, and motor manufacturer Aveox.
  • Last Rounds? Status of Key Munitions at the Iran War Ceasefire — U.S. forces may have expended more than half of prewar inventory for four of seven heavily used munitions in the 39-day air and missile campaign against Iran, with rebuilding to prewar levels projected to take one to four years. These munitions are also critical for a potential Western Pacific conflict.

Thursday, June 25, 2026

AI Summary. An analysis estimates the AI ecosystem generated $110bn in revenue over 12 months excluding double-counting, with the most recent month annualizing to a $175bn run rate. After separating AI-specific capital spending from the roughly $120bn hyperscalers already spent annually and depreciating compute over 6 years, revenues attributable to hyperscalers roughly cover depreciation.

Azeem Azhar, William Gildea, Hannah Petrovic, Nathan Warren, et al. Exponential View
Date Posted:

Depreciation currently absorbs ~81% of the hyperscalers’ revenue and ~68% of gross AI revenue. Depreciating AI compute assets over 6 years, Azeem Azhar argues “revenues attributable to hyperscalers just about clear the depreciation expense.”

Do AI revenues cover the depreciation on AI infrastructure spending?

Core argument: The AI ecosystem generated $110 billion in revenue over the past 12 months after removing double-counting, and the most recent month's revenues annualise to a $175 billion run rate.

Over the past 12 months, the AI ecosystem generated $110 billion in revenue when you remove double-counting. Annualising the most recent month’s revenues indicates a $175 billion revenue run rate. AI revenues cover the capital investment that’s required to build the infrastructure. Our model separates AI-oriented CapEx from ordinary CapEx across the major hyperscalers and neoclouds, the specialist AI cloud providers. This adjustment is important because hyperscalers were already spending around $120 billion annually on CapEx before ChatGPT. We capture the additional investment in AI infrastructure, then depreciate compute assets over 6 years and other infrastructure over 14 years. Our modelling shows that revenues attributable to hyperscalers just about clear the depreciation expense. Six years is defensible. That longer useful life reflects that, one, demand still exceeds available AI compute, and two, operators are getting better at managing GPU fleets. Both help. The second alone is enough to justify a longer economic life.

Takeaways by Macro Roundup® AI

  1. The AI ecosystem generated $110 billion in revenue over the past 12 months after removing double-counting, and the most recent month's revenues annualise to a $175 billion run rate.
  2. The model separates AI-oriented CapEx from ordinary CapEx across major hyperscalers and neoclouds, an adjustment that matters because hyperscalers already spent around $120 billion annually on CapEx before ChatGPT.
  3. Depreciating compute assets over 6 years and other infrastructure over 14 years, the model shows revenues attributable to hyperscalers just about clear the depreciation expense, with the 6-year life supported by demand exceeding available AI compute and improved GPU fleet management.

Related Articles:

  • Semiquincententacles — Leading artificial-intelligence labs are moving to usage-based token pricing on a cost-plus basis that includes total ownership costs, pushing token prices higher and giving some enterprises reason to migrate to cheaper open models. Frontier models remain necessary for cybersecurity, distillation training, agents, and scientific discovery, where maximum performance is required.
  • Why .400 Hitters Disappeared — and What It Means for AI — Artificial intelligence models' capability gains are shrinking in percentage terms, falling from about 15% relative to the baseline early on to roughly 10% for the same absolute increment, narrowing the gap between frontier and good-enough models. The author argues this points to a commoditizing market where price, not minor model differences, becomes the main differentiator, pressuring frontier firms' margins.
  • Four Takeaways from Microsoft’s Satya Nadella at WEF Today — Microsoft's chief executive says the price of AI model usage ("tokens") falls by roughly half every three months. Commentary argues that such rapid deflation pulls demand forward, encourages overbuilding, and commoditizes models, shifting competitive advantage to orchestration and context rather than the model layer.

AI Summary. US communities more exposed to opioid marketing experienced lower population growth, driven mainly by out-migration of college-educated adults, alongside higher fertility among non-college-educated mothers that partially offset the decline. The authors attribute the selective migration response to eroded local quality of life, including addiction, homelessness, and strained public services.

Carolina Arteaga, Victoria Barone, and Stephen Claassen National Bureau of Economic Research
Date Posted:

Commuting zones with high cancer mortality in 1996, and thus more exposed to Purdue’s OxyContin marketing, saw lower population growth via selective out‑migration of college‑educated adults, partially offset by higher fertility among non‑college mothers.

How does opioid exposure reshape local population growth and composition?

Core argument: Using 1996 cancer mortality as a proxy for where pharmaceutical companies initially concentrated OxyContin marketing, the authors find communities more exposed to the opioid epidemic experienced lower population growth, driven primarily by out-migration among college-educated adults.

To measure exposure to opioid shocks, we use 1996 cancer mortality as a proxy for local exposure to the opioid epidemic. The size of the cancer pain market in the mid-1990s was a key determinant of where pharmaceutical companies initially concentrated their marketing efforts for opioids. Using this quasi exogenous variation in local exposure to Purdue Pharma’s marketing of OxyContin, we show that communities more exposed to the epidemic experienced lower population growth, driven primarily by increased out-migration among college-educated adults. In the opposite direction, the opioid epidemic induced substantial increases in fertility, concentrated among non-college educated mothers, which by the end of our period partially offset these population declines. Mortality rises sharply in proportional terms [but] the modest absolute increase has only a minimal effect on population growth. These shifts reveal how the epidemic contributed to growing regional demographic divergence within the United States. Beyond disrupting labor-market opportunities, it eroded the amenity value of places. Communities most exposed to the crisis were characterized by rising addiction, homelessness, neighborhood disorder, and a strain on public services, reducing local quality of life. This deterioration in local amenities helps explain why the migration response we document is both sizable and selective.

Takeaways by Macro Roundup® AI

  1. Using 1996 cancer mortality as a proxy for where pharmaceutical companies initially concentrated OxyContin marketing, the authors find communities more exposed to the opioid epidemic experienced lower population growth, driven primarily by out-migration among college-educated adults.
  2. The opioid epidemic induced substantial fertility increases concentrated among non-college-educated mothers, which by the end of the study period partially offset population declines.
  3. Mortality rose sharply in proportional terms but the modest absolute increase had only minimal effect on population growth, and the authors attribute the sizable, selective migration response to eroded local amenities including rising addiction, homelessness, neighborhood disorder, and strained public services.

Related Articles:

  • Republican Support and Economic Hardship: Enduring Effects of the Opioid Epidemic — The authors estimate that geographic exposure to prescription-opioid marketing raised drug-related mortality and reliance on public transfer programs, and continuously increased the Republican vote share in House, presidential, and gubernatorial elections. They estimate a one-standard-deviation increase in exposure raised the Republican House vote share by 4.5 percentage points.
  • Democracy and the Opioid Epidemic — Using US commuting zone data and the concentration of early prescription opioid marketing in cancer pain markets, the authors estimate that exposure to the opioid epidemic substantially increased the Republican vote share in congressional and presidential elections, with a one standard deviation rise in initial cancer mortality corresponding to 13.8 percentage points.
  • The Rise in American Pain: The Importance of the Great Recession — The authors find that the rise in reported pain among Americans was almost entirely concentrated in the 2007–2010 recession period rather than following a smooth upward trend, with the disproportionate increase among the less educated also occurring primarily then. The authors state they find this jump difficult to explain.

AI Summary. The United States became a net oil exporter in late 2019, and oil and oil product spending fell from near 8% of GDP in 1980 to 3% in 2024. The authors estimate a 15% global oil supply disruption would now cut annualized U.S. real GDP growth by 0.3pp, versus 1.7pp in the rest of the world.

Lutz Kilian, Michael Plante and Alexander Richter Federal Reserve Bank of Dallas
Date Posted:
Is Database:
Database

Kilian, et al find that the impact of an energy shock on US real GDP growth has fallen to 1/20th of what it would have been in 1980, due both to the declining share of oil in GDP and the US transition from net importer to net exporter.

How much do oil supply shocks still hurt US growth?

Core argument: A world oil supply reduction of the size seen in the 2026 Iran War would have cut annualized U.S. real GDP growth by 5.6pp in 1980 but reduces it by only 0.3pp today, one-twentieth of the 1980 decline.

The 2026 disruption of global oil supplies has been more than twice as large as the peak disruption during the 1973 oil crisis, the largest oil supply shortfall on record before 2026. It has been widely noted that the shale oil revolution has allowed the U.S. to greatly reduce its dependence on imports of crude oil and refined products. Whereas the U.S. economy had been a major net importer of oil and oil products since the 1970s, the U.S. became a net exporter in late 2019, if only by a small margin (Chart 2, panel A). In addition, oil and oil product expenditures as a share of GDP substantially declined from a high near 8% in 1980 to 3% in 2024 (Chart 2, panel B). Reduction of world oil supply [as in the 2026 Iran War] would have caused annualized U.S. real GDP growth to decline by 5.6pp in 1980; the same event today reduces growth by only 0.3pp, one-twentieth of the 1980 decline (Chart 3). Our analysis demonstrates that the US contraction in response to a geopolitical oil supply disruption was similar to that in the rest of the world in 1980. This is no longer the case. A global oil supply disruption of 15% today causes a decline in annualized real GDP growth of 1.7pp in the rest of the world, compared with only 0.3pp in the U.S.—roughly one-sixth of the response in the rest of the world.

Takeaways by Macro Roundup® AI

  1. A world oil supply reduction of the size seen in the 2026 Iran War would have cut annualized U.S. real GDP growth by 5.6pp in 1980 but reduces it by only 0.3pp today, one-twentieth of the 1980 decline.
  2. A 15% global oil supply disruption today lowers annualized real GDP growth by 1.7pp in the rest of the world versus 0.3pp in the U.S., roughly one-sixth of the response, whereas the contractions were similar in 1980.
  3. The U.S. became a net exporter of oil and oil products in late 2019, by a small margin, after being a major net importer since the 1970s, and oil and oil product expenditures fell from near 8% of GDP in 1980 to 3% in 2024.

Related Articles:

  • Fighting Words: The Energy Transition in 2026 — Estimates based on primary energy overstate how much electricity is needed to replace oil in passenger cars, since electric motors convert 80%-90% of electricity into motion versus about 18% for gasoline cars. Measuring useful final energy net of waste heat avoids this "primary energy fallacy": the US consumed 45 exajoules of useful final energy versus 95 exajoules of primary energy.
  • 12 Insights About Venezuela And The “Donroe Doctrine” — Petroleum supplies 90% of US transport energy consumption and about 33% of industrial production, mainly high-temperature heat and feedstocks, with passenger cars, SUVs and light trucks accounting for roughly 60% of transport energy use. The oil intensity of GDP is gradually declining in most of the world, though the US retains the highest oil intensity among major economies.
  • Power Politics: Energy Self-Sufficiency in a Modern Mercantilist World — China leads in adding power to the grid, with more than 30,000 miles of ultra-high-voltage transmission lines versus zero in the US and lead times for large transformers of 48 weeks against a US average of 143 weeks. This positions China to accommodate new demand sources such as artificial intelligence.

AI Summary. Since 2017 the Japanese bond market has fallen 66% in value, the German and French markets 35%, and the US market 24%, while China's is the only major bond market up over that period. Few investors in advanced economies hold Chinese government bonds, particularly since Russian reserves were frozen.

Charles Gave Gavekal
Date Posted:

As measured by total real returns in dollars, the only major sovereign 10 year bond that has had a positive return since January 2017 is China’s. The 10-year JGB is down 66%, equivalent French and German debt is down 35%, and the US 10-year has declined 24%.

Have major advanced-economy bond markets lost value while China's gained?

Core argument: The Japanese bond market has lost 66% of its value since January 2017, a steeper decline than the German and French markets at 35% and the US market at 24%.

The Japanese bond market has effectively collapsed in value, down 66% since January 2017, while the German and French bond markets are down 35% and the US bond market is down 24%. The only major bond market that is up since 2017 is China’s. However, it is well understood that few investors in the OECD own CGBs, particularly since Russian reserves were frozen in 2022.

Takeaways by Macro Roundup® AI

  1. The Japanese bond market has lost 66% of its value since January 2017, a steeper decline than the German and French markets at 35% and the US market at 24%.
  2. China's is the only major bond market that is up since 2017.
  3. Few OECD investors hold Chinese government bonds, particularly since the freezing of Russian reserves in 2022.

Related Articles:

  • 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%.
  • Why Global Imbalances Matter — Countries running persistent trade surpluses, such as China, Germany and Japan, have built their economies around an excess supply of tradeable goods and services, leaving them dependent on deficit countries to absorb it. That dependence gives deficit countries bargaining power and exposes surplus countries to pressure to expand domestic demand.
  • The Emergence Of A New Anchor — Renminbi bonds have outperformed other major bond markets over horizons ranging from year-to-date to the past 10 years, and most countries now trade more with China than with the US. The author argues the world is not deglobalizing but "de-Westernizing," with greater trade with China making countries more willing to save in renminbi.

AI Summary. Of roughly 1,900 startups valued at $1bn or more in a Stanford-maintained database, 332 had raised money at or below their peak valuation and 383 had disclosed no new funding in the prior three years. Research firm Pitchbook expects net valuation cuts of $500bn–$1tn as these firms reprice, seek buyers, or fail.

Economist Staff The Economist
Date Posted:
Is Database:
Database

As of May of this year, 17% of unicorns in a Stanford database had raised funds at a valuation at or below their peaks; ~ half the firms in the database were valued under their peak valuation, implying significant write-downs going forward.

How many billion-dollar startups are struggling to justify their valuations?

Core argument: By May 2026, 332 of the 1,900 unicorns in Ilya Strebulaev's Stanford database had raised money at or below their peak valuation, and 212 of those were valued at under $1bn.

By May 2026, 332 of the 1,900 unicorns in a database maintained by Ilya Strebulaev of Stanford University had raised money at a valuation at or below their peak. Of those, 212 were valued at under $1bn. As many as 383 had disclosed no new funding in the previous three years; 41 of these had lost unicorn status. As Mr Strebulaev notes, data on startups are notoriously noisy. Some unicorns may have raised no money because they did not need it. But others may be struggling to justify mythical valuations. It seems that a fair proportion of unicorns have joined the undead. Nearly half of the firms in Mr Strebulaev’s database valued at less than their peak raised money in 2021. From late next year, their investors will start demanding results. Pitchbook expects net cuts in valuations of between $500bn and $1trn as firms reprice, scramble to find a buyer, or go bust. That is quite a cull: Pitchbook puts the total value of unicorns below the ten most valuable at around $5trn.

Takeaways by Macro Roundup® AI

  1. By May 2026, 332 of the 1,900 unicorns in Ilya Strebulaev's Stanford database had raised money at or below their peak valuation, and 212 of those were valued at under $1bn.
  2. As many as 383 unicorns in Strebulaev's database had disclosed no new funding in the previous three years, and 41 of them had lost unicorn status, though Strebulaev notes startup data are notoriously noisy and some firms may not have needed capital.
  3. Pitchbook expects net valuation cuts of between $500bn and $1trn as firms reprice, scramble to find a buyer, or go bust, against roughly $5trn in total value for unicorns below the ten most valuable.

Related Articles:

  • Trampled Underfoot by a Stampede of Unicorns — A database trawl counts 1,727 private companies valued above $1bn globally, and if private market valuations are accurate, the largest of them have outperformed public markets since the pandemic. Several, led by SpaceX and OpenAI, plan initial public offerings, which can signal trouble ahead since insiders are choosing to sell.
  • The Deep End: 2025 Alternative Investments Review — In venture capital, only consistently investing with top-quartile managers outperformed the S&P 500, while the median manager at best matched public equity markets and fourth-quartile managers destroyed value relative to equities. Fees are part of the issue: an 11.7% annualized NASDAQ 100 return falls to 7.4% under a 2%/20% fee structure.
  • Valuing AI: Extreme Bubble, New Golden Era, or Both — By the investment firm's definition of a bubble — a price two standard deviations above an asset class's long-term real trend — U.S. equities are in bubble territory and extremely overpriced on historically effective valuation metrics, with a cyclically adjusted price-earnings ratio of 40 and a record share of the market trading above 10x sales.