Edward Conard

Top Ten New York Times Bestselling Author

  • “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
  • “…a must-read for serious students of economic policy.” - Glenn Hubbard, Dean, Columbia Business School, and former Chairman of the Council of Economic Advisers
  • “…a must-read for serious students of economic policy.” - Glenn Hubbard, Dean, Columbia Business School, and former Chairman of the Council of Economic Advisers
  • “…a fresh argument for the productive value of inequality.” - David Autor, Professor of Economics, Massachusetts Institute of Technology
  • “…challenges misconceptions that distort our economic debates.” - Arthur Brooks, President of the American Enterprise Institute
  • “…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 represents the most cogent and persuasive analysis of the Financial Crisis to date.” - Andrei Shleifer, 1999 John Bates Clark Medal Winner
  • “Unintended Consequences is full of substance, it is one of the must-read books of the year, and once I finish it I will be giving it a second read through right away.” - Tyler Cowen, Professor, George Mason University
  • “…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 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 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
Upside of Inequality Oxford Unintended Consequences

Macro Roundup


Wednesday, July 15, 2026

AI Summary. Using two centuries of U.S. data and a panel of advanced economies, the authors find government primary surpluses are more closely associated with debt-service costs than with debt-to-output ratios, which lose much of their explanatory power once debt service is accounted for. The association is stronger when financing conditions deteriorate.

Barry Eichengreen, Maxime Menuet and Gregory Donnat National Bureau of Economic Research
Date Posted:

Do primary surpluses respond to the stock of debt or to how costly it is to service? Eichengreen et al test this back to 1800. Though the impulse responses are consistent with both, once service costs enter the model, debt-to-GDP loses significance.

Do debt-service costs explain fiscal adjustment better than debt ratios?

Core argument: Using two centuries of U.S. data and a panel of advanced economies, the authors find primary surpluses are more closely associated with debt-service costs than with debt ratios, which lose much of their explanatory power once debt service is included.

Using two centuries of U.S. data and a panel of advanced economies, we show that primary surpluses are more closely associated with debt-service costs than with debt ratios. Once debt service is taken into account, debt ratios lose much of their explanatory power. The evidence also suggests that this relationship is state-dependent. Primary surpluses are more strongly associated with debt-service costs when financing conditions deteriorate, especially when the interest-growth differential [“r-g”] becomes unfavorable. Standard debt-based fiscal reaction functions may overlook an important dimension of fiscal sustainability by focusing on debt accumulation rather than on the budgetary burden it creates. Ultimately, sovereign debt becomes fiscally constraining not when debt ratios reach a particular threshold, but when financing conditions transform outstanding liabilities into an immediate budgetary burden.

Takeaways by Macro Roundup® AI

  1. Using two centuries of U.S. data and a panel of advanced economies, the authors find primary surpluses are more closely associated with debt-service costs than with debt ratios, which lose much of their explanatory power once debt service is included.
  2. The evidence suggests the relationship is state-dependent, with primary surpluses more strongly associated with debt-service costs when financing conditions deteriorate, especially when the interest-growth differential ("r-g") becomes unfavorable.
  3. The authors argue standard debt-based fiscal reaction functions may overlook a dimension of fiscal sustainability, since sovereign debt becomes fiscally constraining when financing conditions turn outstanding liabilities into an immediate budgetary burden rather than at a particular debt-ratio threshold.

Related Articles:

  • Inflation and Debt — Congressional Budget Office long-term projections show sustained primary deficits of 2% of GDP driving the debt-to-GDP ratio steadily higher, a path the agency itself labels unsustainable. Closing a 2% of GDP fiscal gap is economically straightforward for a growing country at peace, but inflation or default would not resolve the problem, since prospective deficits still require continued borrowing.
  • The Budget and Economic Outlook: 2026 to 2036 — The Congressional Budget Office projects the federal deficit will grow from $1.9tn (5.8% of GDP) to $3.1tn (6.7%) over a decade, above the 3.8% averaged over the past 50 years, with rising net interest costs driving much of the increase. Federal debt held by the public is projected to rise from 101% to 120% of GDP.
  • How Might Fiscal Policy Respond to the Rise of Artificial Intelligence? — The Congressional Budget Office estimates that a permanent 0.5 percentage point increase in annual productivity growth would lower publicly held federal debt after 30 years by 39% of GDP, offsetting roughly half the projected rise from 101% of GDP to 175%. Higher revenue and slower spending growth outweigh higher interest costs.

AI Summary. Technology-driven investment booms have historically ended in sharp corrections with wider economic fallout, and the current artificial intelligence build-out is on track to outgrow every previous episode within three years measured from its pre-boom trough. The expansion is increasingly funded by borrowing through bonds, special purpose vehicles and non-bank lenders, alongside circular financing arrangements that create interconnectedness among firms.

Phurichai Rungcharoenkitkul Bank of International Settlements
Date Posted:

A BIS analysis argues the AI buildout is outpacing prior infrastructure booms and warns, “The larger the boom, the deeper the eventual bust.” Their model suggests that investment is 50% above the efficient level.

Is the AI build-out outpacing past technology investment booms?

Core argument: Firms expanding AI capital expenditure have turned increasingly to borrowing through bond issuance, special purpose vehicles, and non-bank lenders, and such leverage aggravates the fragility of rapid investment ramp-ups.

Technological breakthroughs are typically accompanied by investment booms and buoyant macroeconomic activity. Exuberance about the promise of new technologies intensifies competition among firms eager to capture a share of the revenues. The race to get ahead can result in excessive investment that makes the boom unsustainable and prone to disruption. This fragility is further aggravated by the leverage that accompanies the rapid ramp-up of investment. This boom-bust pattern recurs across history, from the US canal mania in the 1830s and the British railway mania in the 1840s to the roaring 20s and the dotcom boom in the late 90s. These episodes all ended in sharp corrections with wider economic fallout. Relative to its pre-boom trough, the current build-out is on track to outgrow every previous episode only three years. The rapid capex expansion has led some to resort increasingly to borrowing, not only through bond issuance but also through special purpose vehicles and non-bank lenders. Some firms have also engaged in so-called circular financing, in which a firm engaging in the build-out takes an equity position in an AI lab in exchange for the lab’s commitment to purchase future compute. Such arrangements have enabled greater deployment of capital, but they also create interconnectedness, shown in panel (b) of Figure 1.

Takeaways by Macro Roundup® AI

  1. Firms expanding AI capital expenditure have turned increasingly to borrowing through bond issuance, special purpose vehicles, and non-bank lenders, and such leverage aggravates the fragility of rapid investment ramp-ups.
  2. Circular financing, in which a firm engaged in the build-out takes an equity position in an AI lab in exchange for the lab's commitment to purchase future compute, has enabled greater capital deployment while creating interconnectedness.

Related Articles:

  • AI Set to be Largest CapEx Cycle Ever … and Soon Majority Externally Financed — Hyperscaler and data-center investment is tracking toward $1.5tn annually by 2030, which in real terms would exceed the 2005-06 housing boom, the 1990s fiber peak, and the 1882 railroad peak combined. External debt and equity are projected to fund close to 90% of the buildout, making it dependent on credit markets.
  • Have We Entered an Era of High Productivity Growth? — Labor productivity has grown solidly while the San Francisco Federal Reserve's total factor productivity index, which adjusts for the equipment and technology workers use, has grown modestly, suggesting gains so far reflect better tools rather than broader advances. A regime-switching model puts the probability of a high-productivity regime at 57% using labor productivity and 21% using total factor productivity.
  • 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.

AI Summary. Data center construction spending rose 23% year over year and made up 8% of private nonresidential construction, while spending on manufacturing buildings—nearly a quarter of that category—fell 22% to a seasonally adjusted annual rate of $174bn, according to the U.S. Census Bureau. Nonresidential construction materials costs are up more than 55% since early 2020.

Bob Tita Wall Street Journal
Date Posted:

Spending on data center construction in May rose 23% y/y, while construction spending on manufacturing buildings fell 22%. The data center buildout could be crowding out non-tech-related fixed investment.

Is data center building the only part of nonresidential construction growing?

Core argument: Data center construction spending rose 23% year-over-year in May and accounted for 8% of total private nonresidential construction spending, according to the U.S. Census Bureau.

Spending on data center construction in May rose 23% from a year earlier, according to the U.S. Census Bureau. Data centers accounted for 8% of the total spending on private, nonresidential construction. Construction spending on manufacturing buildings dropped 22% year-over-year in May to a seasonally adjusted annual rate of $174 billion. Manufacturing makes up the largest part of private, nonresidential construction, accounting for nearly a quarter of the spending in May, according to the Census Bureau. Materials costs for nonresidential construction are up more than 55% since early 2020 and even higher for fabricated steel, copper wire, and some other individual materials, according to the government’s producer price index.

Takeaways by Macro Roundup® AI

  1. Data center construction spending rose 23% year-over-year in May and accounted for 8% of total private nonresidential construction spending, according to the U.S. Census Bureau.
  2. Construction spending on manufacturing buildings fell 22% year-over-year in May to a seasonally adjusted annual rate of $174 billion, while manufacturing remained nearly a quarter of private nonresidential construction spending, according to the U.S. Census Bureau.
  3. Materials costs for nonresidential construction are up more than 55% since early 2020, with larger increases for fabricated steel, copper wire, and some other materials, according to the government's producer price index.

Related Articles:

  • Midyear Outlook — 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.
  • 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.
  • 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.

AI Summary. Public filings indicate AI compute backlogs—contracts for future supply generating no current revenue—exceed $2.2tn, but under reasonable assumptions the net present value of those rental payments after required capital spending equals only about 15% of the backlog. Slowing AI revenue or lost capital-markets access could raise risk premiums on these obligations.

Jason Thomas Carlyle
Date Posted:

AI infrastructure firms have a $2.2T revenue backlog — but it’s unsecured and could evaporate if labs falter. Netting out the capex to deliver it and discounting at 14%, implies equity is thin – just ~15% of the backlog, Carlyle finds.

Are trillions in AI compute backlogs worth much less than advertised?

Core argument: Public filings suggest AI compute backlogs — contracts for future supply that yield no current revenue — now exceed $2.2 trillion.

Public filings suggest that AI compute backlogs [contracts for future supply that yields no current revenue] now exceed $2.2 trillion. For AI infrastructure providers, these contracts represent both an asset (the discounted present value of compute rental payments from AI labs) and a liability (the discounted present value of the hardware purchases, data center rent, and utility outlays necessary to deliver the promised compute). The implied "equity" in the infrastructure layer looks thin: under reasonable assumptions, the net present value of the rental payments, after the capex required to deliver them, equals only about 15% of the backlog. A deceleration in AI revenue, a derating of the labs, or a loss of capital-markets access could raise the risk premium applied to these longer-dated, unsecured cash flows enough to make these “remaining performance obligations” more trouble than they're worth, imperiling demand for memory chips.

Takeaways by Macro Roundup® AI

  1. Public filings suggest AI compute backlogs — contracts for future supply that yield no current revenue — now exceed $2.2 trillion.
  2. Under reasonable assumptions, the net present value of AI compute rental payments, after the capex required to deliver them, equals only about 15% of the backlog, leaving thin implied equity in the infrastructure layer.
  3. A deceleration in AI revenue, a derating of the labs, or a loss of capital-markets access could raise the risk premium on these longer-dated, unsecured cash flows enough to imperil demand for memory chips.

Related Articles:

  • A Slower AI Payoff Would Be Everyone’s Problem — If artificial intelligence profits arrive later than expected, committed capital spending and depreciation would still hit on schedule, squeezing margins, stretching balance sheets toward more debt and possible ratings downgrades, and repricing equities. Because a handful of large technology firms dominate the indices, the author argues the damage would spread to chips, power, data centers and the broader market.
  • The State of the AI Economy — 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.
  • 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.

AI Summary. Investor demand for hyperscaler bond deals has weakened, with the cover ratio — dollars of investor orders per dollar of bonds issued — falling from nearly 5x to below 2x. This suggests investors may require wider spreads to absorb additional hyperscaler bond supply.

Torsten Sløk Apollo
Date Posted:

The cover ratio for hyperscaler debt issuance has declined from 5x in February 2026 to < 2x in July. Sløk notes yields may need to rise, “suggesting investors may need wider spreads to absorb additional hyperscaler supply.”

Is investor demand for hyperscaler bonds weakening?

Core argument: Cover ratios on hyperscaler bond deals, which measure dollars of investor orders per dollar of bonds issued, fell from nearly 5x in February 2026 to below 2x in July.

The cover ratio measures how many dollars of investor orders a bond deal receives for every dollar of bonds issued. For hyperscalers, it has fallen from nearly 5x in February 2026 to below 2x in July, suggesting investors may need wider spreads to absorb additional hyperscaler supply.

Takeaways by Macro Roundup® AI

  1. Cover ratios on hyperscaler bond deals, which measure dollars of investor orders per dollar of bonds issued, fell from nearly 5x in February 2026 to below 2x in July.
  2. The decline in hyperscaler cover ratios suggests investors may require wider spreads to absorb additional hyperscaler bond supply.

Related Articles:

  • AI Set to be Largest CapEx Cycle Ever … and Soon Majority Externally Financed — Hyperscaler and data-center investment is tracking toward $1.5tn annually by 2030, which in real terms would exceed the 2005-06 housing boom, the 1990s fiber peak, and the 1882 railroad peak combined. External debt and equity are projected to fund close to 90% of the buildout, making it dependent on credit markets.
  • Bond Traders Stunned as Losses on SpaceX’s New Debt Keep Growing — SpaceX's $25bn bond offering has weakened in secondary trading, with paper losses relative to Treasuries totaling roughly $305m. Traders attribute the selling to fast-money accounts seeking quick profits rather than buy-and-hold investors, though the company also won investment-grade ratings despite expectations of years of negative cash flow.
  • A Slower AI Payoff Would Be Everyone’s Problem — If artificial intelligence profits arrive later than expected, committed capital spending and depreciation would still hit on schedule, squeezing margins, stretching balance sheets toward more debt and possible ratings downgrades, and repricing equities. Because a handful of large technology firms dominate the indices, the author argues the damage would spread to chips, power, data centers and the broader market.

AI Summary. Sub-Saharan Africa's cereal output is nearly 5x its 1960s level, but most gains came from cultivating more land rather than higher yields, and arable land per person has been falling for decades. Food and Agriculture Organisation data show cereal yields flat in recent years, with most African countries recording lower agricultural total factor productivity than a decade earlier.

Economist Staff The Economist
Date Posted:
Is Database:
Database

Most Sub-Saharan African countries had lower agricultural TFP in 2023 than they had a decade earlier. Btw 2008 and 2019, in the face of a rapidly rising population, small farms’ yields were falling by 3 to 4% per year.

Is Africa's crop growth coming from more land, not higher yields?

Core argument: Sub-Saharan Africa's farmers grow nearly five times as much cereal, such as maize and rice, as in the 1960s, but most of those gains came from cultivating more land rather than raising output per hectare.

In aggregate [sub-Saharan Africa's] farmers are growing more cereals, such as maize (corn) and rice, than ever: nearly five times as much as in the 1960s. But most of those gains came from cultivating more land, which cannot go on for ever. The amount of arable land per person has been falling for decades, and now sits at roughly the global average. That might not matter if farmers were also growing more crops per hectare. But recently gentle growth in agricultural productivity has given way to stagnation, perhaps even decline. Consider figures drawn from national statistics in Africa by the Food and Agriculture Organisation. Cereal yields did not grow between 2020 and 2024, the latest data point. Nor did total factor productivity (TFP), a measure of how efficiently inputs of all kinds (such as labour and machinery) are turned into produce. Most African countries had lower agricultural TFP in 2023 than a decade before.

Takeaways by Macro Roundup® AI

  1. Sub-Saharan Africa's farmers grow nearly five times as much cereal, such as maize and rice, as in the 1960s, but most of those gains came from cultivating more land rather than raising output per hectare.
  2. Cereal yields in Africa did not grow between 2020 and 2024, and neither did total factor productivity, a measure of how efficiently inputs such as labour and machinery are turned into produce, according to Food and Agriculture Organisation figures drawn from national statistics.
  3. Arable land per person in sub-Saharan Africa has been falling for decades and now sits at roughly the global average, while most African countries had lower agricultural total factor productivity in 2023 than a decade before.

Related Articles:

  • 21st-Century Africa Governance and Growth — Sub-Saharan Africa's GDP per capita has fallen relative to the world average over three decades, from 30% to 22%, reflecting the region's inability to sustain growth. Had it grown at the global pace since 1990, income per capita would be over 40% higher; at emerging East Asia's pace, nearly 3x higher.
  • Africa Needs 1 Billion Jobs by the End of the Century. Where Will It Find Them? — Africa's population, now about 1.5bn, is projected to reach 4bn by 2100, with sub-Saharan Africa approaching 3.5bn and Nigeria's working-age population forecast to exceed 320m. Between 2005 and 2020, Congo, Nigeria, Ethiopia and other large states averaged only one job created for every two entrants to the working-age population.
  • Changing Sub-Saharan Africa’s Growth Trajectory — Sub-Saharan Africa's working-age population is growing nearly 3% a year, versus close to zero in much of the world, and is projected to exceed China's within about a decade and reach a quarter of the global total. With modest productivity growth near 2%, the region is likely among the world's fastest-growing.

Thursday, July 16, 2026

AI Summary. The headline labor force participation rate — the share of people 16 and older working or looking for work — is a limited gauge of labor market health because participation varies greatly by age. Aging baby boomers and the addition of nearly 1.7m Americans 65 and older to population estimates add distortion.

Justin Fox Bloomberg
Date Posted:
Is Database:
Database
Is Important:
Important

Adjusted for the age composition of the population, US labor force participation, at 61.8%, is near an all-time high, reflecting delayed retirement of older workers.

Does the headline labor force participation rate accurately measure labor market health?

Core argument: The Bureau of Labor Statistics added nearly 1.7 million Americans aged 65 and older to its population estimates this year, reflecting belated adjustments stemming from the 2020 Census, which adds to distortions in the labor force participation rate.

The labor force participation rate is the number of people 16 and older who have jobs or are actively looking for them, divided by the civilian (that is, not uniformed military) noninstitutional (that is, not in prison or the like) population 16 and older. While this percentage can tell us useful things about, say, the sustainability of Social Security or the overall availability of labor, its validity as a gauge of the health of the labor market is limited by the reality that labor force participation varies greatly by age. It’s a flawed metric that is currently more flawed than usual as the youngest members of the giant baby boom generation age into their mid-60s. The addition this year of nearly 1.7 million 65-and-older Americans to the BLS population estimates, because of belated adjustments stemming from the 2020 Census, just adds to the distortions.

Takeaways by Macro Roundup® AI

  1. The Bureau of Labor Statistics added nearly 1.7 million Americans aged 65 and older to its population estimates this year, reflecting belated adjustments stemming from the 2020 Census, which adds to distortions in the labor force participation rate.
  2. The labor force participation rate divides people 16 and older who are working or actively looking for work by the civilian noninstitutional population 16 and older, a measure whose validity as a gauge of labor market health is limited because participation varies greatly by age.
  3. The aging of the youngest baby boomers into their mid-60s makes the participation rate more flawed than usual, though the measure still tells us useful things about Social Security sustainability and the overall availability of labor.

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.
  • AI, Immigration, and Collapsing Labor Force Participation — As labor force participation falls, a shrinking group of workers generates the income, taxes, and innovation supporting a growing number of non-workers, leaving the economy more fragile because labor shocks are concentrated among fewer participants. Labor scarcity raises the return on automation, robotics, and software, making growth depend on productivity rather than participation.
  • How Labor Force Participation Has Diverged Across Genders — Between 1976 and 2024 the trend labor force participation rate fell 10.1 percentage points for men and rose 8.2 points for women. Population aging lowered participation by about 4 points for each gender and rising education raised it (+6 for women, +3 for men), but the divergence is dominated by participation trends within age-gender groups.

AI Summary. The authors find no statistically significant evidence that $190bn in federal emergency K-12 school aid mitigated pandemic learning loss or raised district spending. Districts qualifying for additional funds saw local revenues fall by $907 per pupil over three years, about 8% of pre-pandemic average, roughly 80% from property taxes.

Jeffrey Clemens, Philip Hoxie and Stan Veuger National Bureau of Economic Research
Date Posted:
Is Database:
Database
Is Important:
Important

Exploiting a discontinuity in K-12 public school districts’ qualification for Covid emergency funding (ESSR), Clemens et al find that the funds did not mitigate learning loss. To a large extent, local districts used them to lower property taxes.

Do federal emergency education grants reduce student learning loss?

Core argument: Districts crossing the 5% poverty cutoff for additional ESSER funds saw local revenues fall an estimated $907 per pupil over three years, about 8% of the pre-COVID average.

Emergency funding for K-12 public schools was an important component of recession spending packages during COVID, involving $190 billion in additional expenditure. We see no evidence that ESSER funds helped to mitigate learning loss, at least in the short run. We find no statistically significant evidence that districts increased expenditures in SY 2021 or SY 2022. We do find evidence that districts that qualified for additional ESSER funds had statistically significant reductions in local revenues that were in excess of their ESSER funds. This pass-through of federal funds may partially explain why ESSER had minimal impacts on learning loss for districts around the 5% poverty threshold. Our estimates suggest that over the three years, crossing the qualification cutoff for additional ESSER funds is associated with a $907 per pupil decrease in local revenues, or about 8% of the pre-COVID average. We show the decline in local revenues as an event study in Figure 3, which includes a flat pre-COVID trend in local revenues per pupil across the qualification threshold. In Table 5,Column 5 [], we can see that about 80% of the revenue decline came from decreases in revenues from property taxes, [which fell] by $718 per pupil across the 3 years. The findings apply specifically to districts in the neighborhood of the 5% poverty threshold for qualifying for additional ESSER funds.

Takeaways by Macro Roundup® AI

  1. Districts crossing the 5% poverty cutoff for additional ESSER funds saw local revenues fall an estimated $907 per pupil over three years, about 8% of the pre-COVID average.
  2. About 80% of that local revenue decline came from property taxes, which fell by $718 per pupil across the three years.
  3. The authors find no evidence that ESSER emergency K-12 funding, totaling $190 billion, mitigated learning loss in the short run, and no statistically significant increase in district expenditures in SY 2021 or SY 2022.

Related Articles:

AI Summary. The author argues that a surge in money growth beginning in 2020 put upward pressure on inflation to a degree unprecedented in the sample period since 1967, and that subsequent monetary tightening produced strong disinflationary pressures. Had policymakers tracked money growth using a price-target model, they might have tightened sooner and avoided some of the inflation.

Peter Ireland, Stephen Miran and Nouriel Roubini Hudson Bay Capital
Date Posted:
Is Database:
Database
Is Important:
Important

Ireland, Miran, and Roubini compare the actual price level to the predictions of an equilibrium model relating prices to money supply. Predicted inflation hit a 60-year high in 2020–21, months before inflation surged and then crashed once the Fed hiked.

Would tracking money growth have helped policymakers tighten sooner?

Core argument: The money growth surge beginning in 2020 and continuing into 2021 put upward pressure on inflation to a degree unprecedented in the post-1967 sample period, according to the graphs in Figure 3.

The graphs in Figure 3 show quite clearly how the surge in money growth starting in 2020 and continuing in 2021 put enormous upward pressure on inflation, to a degree unprecedented in the post-1967 sample period. And while the large and negative price gaps that followed in 2022 and 2023 are likewise indicative of strong disinflationary pressures applied through subsequent monetary tightening, one can’t see these graphs without asking: Had Federal Open Market Committee members been monitoring measures of money growth with the help of the P-star [price target based on monetary aggregates] model, might they have ended QE and raised interest rates sooner and more quickly, thereby avoiding at least some of the post-2020 inflation? Of course, massive fiscal expansion was another driver of the post-2020 surge in inflation, as suggested by fiscal theories of the price level. The model simply observes that regardless of its originating source, excessive money growth signals that inappropriate macroeconomic policies are fueling higher inflation. Fiscal expansions that expand money supply will be reflected in a P-star model.

Takeaways by Macro Roundup® AI

  1. The money growth surge beginning in 2020 and continuing into 2021 put upward pressure on inflation to a degree unprecedented in the post-1967 sample period, according to the graphs in Figure 3.
  2. The large negative price gaps of 2022 and 2023 indicate strong disinflationary pressures from subsequent monetary tightening, raising the question of whether FOMC members using the P-star model would have ended quantitative easing and raised interest rates sooner and more quickly.
  3. Massive fiscal expansion was another driver of the post-2020 inflation surge, as suggested by fiscal theories of the price level, and fiscal expansions that expand the money supply are reflected in a P-star model.

Related Articles:

  • Money and Inflation — The M2 money supply has contracted for the first time in sixty years, shrinking at an annual rate of nearly 5%, yet inflation has remained high rather than turning to deflation. The author argues this undercuts the claim that money supply growth drives inflation, since the same logic implies severe deflation should now follow.
  • State Dependence of Monetary Policy During Global Supply Chain Disruptions — The authors present descriptive evidence that global supply chain disruptions were associated with elevated transportation costs, goods supply-demand imbalances, and a surge in goods prices, steepening the aggregate supply curve. In the disrupted state, a contractionary monetary policy shock produces muted responses in real output and spare capacity but a substantially stronger decline in goods prices.
  • What Next for r*? A Capital Market Equilibrium Perspective On The Natural Rate of Interest — A capital-market equilibrium model estimates that the steady-state safe natural rate of interest across advanced economies has fallen from about 5% to about 0%, driven by slower growth, demographics, higher mark-ups, and rising risk premia. The author estimates that artificial-intelligence-driven growth and heightened inflation risk could each raise it by about 1 percentage point.

AI Summary. US inflation-protected Treasury yields have broken above their post-2008 highs, indicating the real cost of capital is rising, while nominal 20- and 30-year Treasury yields remain within their three-year ranges. The article argues an artificial-intelligence-driven investment boom has raised demand for capital, and that corporate debt-to-equity ratios have fallen and interest coverage has surged since Covid.

Tan Kai Xian Gavekal
Date Posted:

Tan Kai Xian argues the rising cost of capital in the US is unlikely to be the reason that non-AI CAPEX is as low as it is, as “healthy profitability has allowed businesses outside the tech hyperscaler sector to fund their capex internally.”

Are rising real yields a sign of an investment boom?

Core argument: Yields on long-term US Treasury inflation-protected securities have broken above their post-2008-crisis highs, while nominal 20-year and 30-year Treasury yields remain within their ranges of the past three years, indicating the real cost of capital in the US is rising.

US real yields are rising. Long-term nominal yields are trending higher, too. But nominal yields on 20-year and 30-year treasuries are still within their ranges of the past three years. In contrast, the yields on long-term US treasury inflation-protected securities have broken above their post-2008-crisis highs. This means the real cost of capital in the US is rising. AI optimism has led to an investment boom fueled by capital that remains relatively cheap compared with the returns on capital. The resulting demand for capital has seen the cost of capital begin to rise. This makes sense in an economic upcycle. And the dynamic could continue for a good while yet; the US is likely nearer the beginning of this growth phase than the end. Corporate debt-to-equity ratios have fallen since Covid, interest coverage ratios have surged, and healthy profitability has allowed businesses outside the tech hyperscaler sector to fund their capex internally.

Takeaways by Macro Roundup® AI

  1. Yields on long-term US Treasury inflation-protected securities have broken above their post-2008-crisis highs, while nominal 20-year and 30-year Treasury yields remain within their ranges of the past three years, indicating the real cost of capital in the US is rising.
  2. AI optimism has driven an investment boom funded by capital still cheap relative to returns on capital, and the resulting demand for capital has begun pushing the cost of capital higher, consistent with an economic upcycle.
  3. Corporate debt-to-equity ratios have fallen since Covid and interest coverage ratios have surged, with healthy profitability allowing businesses outside the tech hyperscaler sector to fund capital expenditure internally.

Related Articles:

  • Capital Allocation — Internal financing has been the main source of capital for U.S. corporations, with steady new debt issuance except in recession years, consistent with a stable capital structure. Companies have retired equity on average, and net equity issuance remains negative even after accounting for stock-based compensation.
  • Cover Ratios for Hyperscaler Bonds Declining — Investor demand for hyperscaler bond deals has weakened, with the cover ratio — dollars of investor orders per dollar of bonds issued — falling from nearly 5x to below 2x. This suggests investors may require wider spreads to absorb additional hyperscaler bond supply.
  • US Stock Market To Stop Shrinking For First Time In 23 Years — Goldman Sachs estimates that net supply of US equity — new shares issued less shares removed by buybacks and take-privates — will be roughly flat in 2026 after being negative since 2003, with a larger influx expected in 2027 as lock-up periods on this year's initial public offerings expire.

AI Summary. The present value of future rental payments from the roughly $2.2tn artificial intelligence compute backlog, after deducting the capital spending needed to deliver that compute, equals only about 15% of the backlog. That thin implied-equity layer means the contracts hold value only if customers stay creditworthy, capital remains available, and the infrastructure earns competitive returns for years.

Paul Kedrosky Applied Complexity
Date Posted:

Kedrosky notes the AI revenue backlog is not a large safety margin: “Slower AI growth, weaker lab valuations, or restricted capital-market access could sharply reduce the present value of long-dated, generally unsecured commitments.”

How much economic value underlies the AI compute backlog?

Core argument: The net present value of future AI compute rental payments, after deducting the capex needed to provide that compute, equals only about 15% of the roughly $2.2tn hyperscaler capital expenditure backlog.

Remarkably little economic value sits beneath the massive AI capital expenditure backlog. The net present value of future rental payments—after deducting the capex required to provide the compute—equals only about 15% of the roughly $2.2 trillion backlog. That thin implied-equity layer means the headline contracts are valuable only so long as customers remain creditworthy, capital stays available, and the infrastructure earns competitive returns over many years. That is heroic stuff at this scale.

Takeaways by Macro Roundup® AI

  1. The net present value of future AI compute rental payments, after deducting the capex needed to provide that compute, equals only about 15% of the roughly $2.2tn hyperscaler capital expenditure backlog.
  2. The thin implied-equity layer means headline AI compute contracts hold value only if customers remain creditworthy, capital stays available, and the infrastructure earns competitive returns over many years.

Related Articles:

  • More Money, More Problems — Public filings indicate AI compute backlogs—contracts for future supply generating no current revenue—exceed $2.2tn, but under reasonable assumptions the net present value of those rental payments after required capital spending equals only about 15% of the backlog. Slowing AI revenue or lost capital-markets access could raise risk premiums on these obligations.
  • AI Set to be Largest CapEx Cycle Ever … and Soon Majority Externally Financed — Hyperscaler and data-center investment is tracking toward $1.5tn annually by 2030, which in real terms would exceed the 2005-06 housing boom, the 1990s fiber peak, and the 1882 railroad peak combined. External debt and equity are projected to fund close to 90% of the buildout, making it dependent on credit markets.
  • Cover Ratios for Hyperscaler Bonds Declining — Investor demand for hyperscaler bond deals has weakened, with the cover ratio — dollars of investor orders per dollar of bonds issued — falling from nearly 5x to below 2x. This suggests investors may require wider spreads to absorb additional hyperscaler bond supply.

AI Summary. Black and Hispanic students make up about 62% of New York City public school enrollment but about 10% of the incoming freshman class at the city's eight most prestigious high schools, where roughly 80% of seats went to Asian and white students. Asian students are about 19% of public school enrollment and received 57% of specialized-school offers.

Matthew Haag New York Times
Date Posted:

62% of New York City public school students are black or Hispanic, and 19% are Asian. Stuyvesant, an elite public school, offered admission to 777 students. Three offers went to black students, and 21 to Hispanics; 69% went to Asian students.

How do admissions at elite public high schools compare with district enrollment?

Core argument: Stuyvesant High School made 3 of its 777 admissions offers to Black students and 21 to Hispanic students, together a nearly one-third decline from the previous freshman class.

About 62% of the students in the city’s public schools are Black or Hispanic. At [NYC] eight most prestigious high schools, about 10% of the students in the incoming freshman class are expected to be Black or Hispanic, roughly the same as last year. About 80% of the seats went to Asian and white students. At Stuyvesant High School in Manhattan, often considered the most selective of the city’s specialized high schools, three of the 777 offers were made to Black students, and 21 were for Hispanic students, together representing a nearly one-third decline from the previous freshman class. About 19% of all public school students are Asian, but they received 57% of the offers for the specialized schools for the fall. The rates were higher at some schools, including 69% at Stuyvesant and 78% at Queens High School for the Sciences at York College.

Takeaways by Macro Roundup® AI

  1. Stuyvesant High School made 3 of its 777 admissions offers to Black students and 21 to Hispanic students, together a nearly one-third decline from the previous freshman class.
  2. Black and Hispanic students make up about 62% of New York City public school enrollment but about 10% of the incoming freshman class at the city's eight most prestigious high schools, roughly unchanged from last year.
  3. Asian students account for about 19% of New York City public school students but received 57% of specialized high school offers for the fall, including 69% at Stuyvesant and 78% at Queens High School for the Sciences at York College.

Related Articles:

  • White Flight from Asian Immigration: Evidence from California Public Schools — In California public schools, the authors estimate that the arrival of one Asian student in a high-socioeconomic-status suburban district is associated with the departure of 1.5 white students, alongside rising academic performance. They suggest this reflects parental concerns about academic competition rather than housing market dynamics or direct racial animus.
  • 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.
  • Americans and Affirmative Action: How the Public Sees the Consideration of Race in College Admissions, Hiring — Survey research finds more Americans disapprove than approve of colleges considering race and ethnicity in admissions: half of U.S. adults said they disapprove of selective colleges weighing race and ethnicity to increase diversity, a third approved, and 16% were unsure. In another survey, 82% said colleges should not consider race or ethnicity, versus 17% who said they should.