Edward Conard

Top Ten New York Times Bestselling Author

  • “…reminds us that inequality sends a signal of what society lacks most, in America’s case, entrepreneurship and risk taking.” - Lawrence Lindsey, CEO, The Lindsey Group, former Director of the National Economic Council
  • “…a must-read for serious students of economic policy.” - Glenn Hubbard, Dean, Columbia Business School, and former Chairman of the Council of Economic Advisers
  • “…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 very valuable contribution.” - Larry Summers, former Secretary of the Treasury and director of the National Economic Council, president emeritus, Harvard 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
  • “…a fresh argument for the productive value of inequality.” - David Autor, Professor of Economics, Massachusetts Institute of Technology
  • “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
  • “…a fresh argument for the productive value of inequality.” - David Autor, Professor of Economics, Massachusetts Institute of Technology
  • “…a very valuable contribution.” - Larry Summers, former Secretary of the Treasury and director of the National Economic Council, president emeritus, Harvard University
  • “…a comprehensive explanation of the modern economy.” - Julian Robertson, Founder, Tiger Management
Upside of Inequality Oxford Unintended Consequences

Macro Roundup


Thursday, March 28, 2024

AI Summary. The authors model automation of human tasks and find that if the complexity of human tasks is unbounded with a sufficiently thick tail, wages can rise indefinitely, but with a thinner tail automation outpaces capital accumulation and wages collapse. If task complexity is bounded, wages first surge, then collapse.

Anton Korinek and Donghyun Suh National Bureau of Economic Research
Date Posted:
Is Database:
Database

.@akorinek et al discuss two scenarios where AI leads to a wage collapse: if AI eventually automates all human tasks, or if capital investment fails to support automation that could potentially augment human productivity.

Does task complexity determine whether automation raises or collapses wages?

Core argument: Under a bounded distribution of task complexity, reflecting the finite computational capacity of the human brain, full automation arrives within finite time once the frontier of automation crosses the maximum complexity of tasks humans perform.

First, we consider the possibility that human tasks are of unbounded complexity [left-hand panel of Figure 1]. In this case, advances in the automation index imply that more and more tasks are automated over time, but that there always remain tasks that cannot be automated. Second, we consider a bounded distribution of task complexity, which reflects that the computational capabilities of the human brain are finite. Bounded distributions result in full automation within finite time when the frontier of automation crosses the maximum complexity of tasks performed by humans. [In the first case] we show that wages can rise indefinitely if the tail is sufficiently thick, as capital accumulation automates ever more complex tasks but there always remains enough for human labor. However, if the tail is too thin, then automation ultimately outpaces capital accumulation and causes a collapse in wages. [In the second case,] we demonstrate that wages would at first surge as machines displace more and more human labor, but would eventually collapse, even before full AGI is reached.

Takeaways by Macro Roundup® AI

  1. Under a bounded distribution of task complexity, reflecting the finite computational capacity of the human brain, full automation arrives within finite time once the frontier of automation crosses the maximum complexity of tasks humans perform.
  2. In the bounded case, wages would at first surge as machines displace more and more human labor, but would eventually collapse, even before artificial general intelligence is reached.
  3. If human tasks are of unbounded complexity, wages can rise indefinitely when the tail of the complexity distribution is sufficiently thick, but collapse if the tail is too thin and automation outpaces capital accumulation.

Related Articles:

  • Assessing the Implications of a Productivity Miracle — If artificial intelligence can perform economically valuable cognitive tasks at very low marginal cost, consumers would buy far more of the cheapened goods and spend the savings in other categories, expanding activity and employment in sectors facing new demand. Where supply is inherently limited, such as real estate, those savings could bid up prices and produce inflation.
  • Generative AI at Work — A study of 5,179 customer support agents finds that access to a generative AI conversational assistant raised issues resolved per hour by 14% on average, with the largest gains for novice and low-skilled workers and minimal effect on experienced, highly skilled ones. The authors also find AI assistance improved customer sentiment, reduced requests for managerial intervention, and improved employee retention.
  • The Short-Term Effects of Generative Artificial Intelligence on Employment: Evidence from an Online Labor Market — Freelancers in occupations most exposed to generative AI saw employment and earnings decline, with the authors estimating a 2% drop in jobs and a 5.2% drop in monthly earnings on an online labor platform. Offering high-quality service did not mitigate the effect, and top-rated freelancers appear disproportionately hurt.

AI Summary. Amazon has committed roughly $148bn over two years to build and operate data centers worldwide, expanding hubs in northern Virginia and Oregon while entering new locations including Mississippi, Saudi Arabia, and Malaysia. The buildout targets demand for corporate services like file storage and databases, as well as the computing power for expected growth in generative artificial intelligence.

Matt Day Bloomberg
Date Posted:

Amazon is planning on spending at least $148B over the next 15 years to build and operate data centers. @markets

Is Amazon spending nearly $150 billion on global data centers?

Core argument: Amazon has committed $148 billion to build and operate data centers worldwide, according to a Bloomberg tally.

Over the past two years, according to a Bloomberg tally, Amazon has committed to spending $148 billion to build and operate data centers around the world. The company plans to expand existing server farm hubs in northern Virginia and Oregon as well as push into new precincts, including Mississippi, Saudi Arabia, and Malaysia. Much of Amazon’s data center expansion is geared toward meeting a rise in demand for corporate services like file storage and databases. But the facilities, along with advanced and expensive chips, will also provide the massive computing power required for an expected boom in generative artificial intelligence.

Takeaways by Macro Roundup® AI

  1. Amazon has committed $148 billion to build and operate data centers worldwide, according to a Bloomberg tally.
  2. Amazon plans to expand existing server farm hubs in northern Virginia and Oregon while pushing into new precincts including Mississippi, Saudi Arabia, and Malaysia.
  3. Much of Amazon's data center expansion targets rising demand for corporate services such as file storage and databases, with the facilities and advanced, expensive chips also supplying computing power for an expected generative AI boom.

Related Articles:

  • AI Datacenter Energy Dilemma – Race for AI Datacenter Space — Artificial intelligence is projected to push datacenters to 4.5% of global electricity generation by 2030, with datacenter power capacity growth accelerating from a 12-15% annual rate to 25%. Global datacenter critical IT power demand is projected to rise from 49 gigawatts to 96 gigawatts, about 40 gigawatts of it for AI.
  • Microsoft Infrastructure – AI & CPU Custom Silicon Maia 100, Athena, Cobalt 100 — Microsoft has planned annual datacenter spending of more than $50bn, which exceeds the annual spend of megaprojects such as nationwide rail networks, dams, and the Apollo moon program. The buildout is aimed at accelerating artificial general intelligence and extending generative AI from productivity applications to leisure.
  • Electricity 2024 — More than 8,000 data centers operate globally, with about 33% in the United States, 16% in Europe, and close to 10% in China. US data center electricity use is projected to rise from roughly 200 TWh in 2022, about 4% of US electricity demand, to almost 260 TWh, or 6% of total demand.

AI Summary. Immigrants accounted for 36% of new U.S. business launches last year, up from 25% in 2019, driven by a rise in Latino entrepreneurship, while business creation by white and native-born Americans has slowed. Hispanic-owned startups tend to be smaller, with 46% reporting over $100,000 in annual revenue versus 62% of white-owned firms.

Ruth Simon Wall Street Journal
Date Posted:

Immigrants from Latin America accounted for 36% of new firm creation in 2023 and started businesses at twice the rate of the U.S. population as a whole. @WSJ

Are immigrants accounting for a growing share of new U.S. businesses?

Core argument: Immigrants accounted for 36% of new U.S. business launches last year, up from 25% in 2019, a rise driven by increased Latino entrepreneurship, according to an analysis of Census Bureau data.

The jump in Latino entrepreneurship has driven up the overall share of new businesses owned by immigrants, who accounted for 36% of launches last year compared with 25% in 2019, according to a new analysis of Census Bureau data. New-business creation by white and native-born Americans has slowed in the past two years, following a broad surge early in the pandemic. Hispanic startups tend to be smaller according to the 2022 Federal Reserve Small Business Credit Survey, with 46% reporting more than $100,000 in annual revenue compared with 62% of white-owned firms, according to the analysis, which looked at firms in operation for five years or less.

Takeaways by Macro Roundup® AI

  1. Immigrants accounted for 36% of new U.S. business launches last year, up from 25% in 2019, a rise driven by increased Latino entrepreneurship, according to an analysis of Census Bureau data.
  2. New-business creation by white and native-born Americans has slowed over the past two years, following a broad surge early in the pandemic.
  3. Among firms operating five years or less, 46% of Hispanic-owned startups reported more than $100,000 in annual revenue, compared with 62% of white-owned firms, according to the 2022 Federal Reserve Small Business Credit Survey.

Related Articles:

  • America’s New Business Boom — The share of US workers employed by young firms is rising for the first time in decades, with 11.3% of jobs at firms less than 5 years old and 8.6% at firms 6–10 years old. Gains in the share of jobs at firms under 5 years old were larger in transportation and warehousing, management and information, and manufacturing.
  • Business Applications Eked Out a New Record in 2023 — Americans filed 5.5m applications to start new businesses in a single year, nearly 1.8m of them highly likely to hire employees — both series highs in U.S. Census Bureau Business Formation Statistics. Likely-employer applications ran 37% above pre-pandemic levels, evidence that startup activity has settled at a higher baseline.
  • Creative Destruction After the Pandemic — US establishment closures spiked and then returned to normal levels, while new establishment openings reached record highs—peaking at 379,000 in a single quarter—and have since cooled but remain far above any pre-2020 level. Much of this may be new businesses replacing those that closed, and it is unclear whether the new entrants are more productive than the firms they replaced.

AI Summary. South Korea's total fertility rate fell from 0.78 to 0.72 and is projected to reach 0.68, far below the 2.1 the OECD says is needed for a stable population. Military personnel numbers have dropped below 500,000, schools and universities report shrinking enrollment, and kindergartens have been converted into nursing homes.

Christian Davies, Song Jung-a and Kang Buseong Financial Times
Date Posted:

South Korea’s record low fertility rate is impacting multiple sectors: “military personnel fell below 500,000 for the first time, universities and schools reported dwindling student numbers, and kindergartens have been converted into nursing homes.” @FT

Are cash incentives enough to reverse a falling fertility rate?

Core argument: South Korea's total fertility rate fell from 0.78 in 2022 to 0.72 in 2023 according to government figures, and is projected to reach 0.68 this year, far below the 2.1 the OECD says is needed for a broadly stable population.

South Korea’s total fertility rate — the average number of children a woman is expected to give birth to in her lifetime — fell from 0.78 in 2022 to 0.72 in 2023, according to government figures. It is projected to fall to 0.68 this year, far below the 2.1 the OECD says is necessary to ensure a broadly stable population. The effects of the crisis are already being felt. In 2022, the number of South Korean military personnel fell below 500,000 for the first time, while universities and schools reported dwindling student numbers and kindergartens have been converted into nursing homes. Last year, more baby carriages were sold in South Korea for pets than for babies. “What people are experiencing in South Korea now is not the result of the 0.7 fertility rate, but of when the fertility rate was 1.2,” said Choi Seul-gi, a researcher at the Korea Development Institute. “It is going to get much worse.”

Takeaways by Macro Roundup® AI

  1. South Korea's total fertility rate fell from 0.78 in 2022 to 0.72 in 2023 according to government figures, and is projected to reach 0.68 this year, far below the 2.1 the OECD says is needed for a broadly stable population.
  2. South Korean military personnel fell below 500,000 for the first time in 2022, while universities and schools reported dwindling student numbers and kindergartens were converted into nursing homes.
  3. Korea Development Institute researcher Choi Seul-gi argues South Korea's present strains reflect the era when fertility was 1.2 rather than 0.7, and that conditions are going to get much worse.

Related Articles:

  • South Korea Keeps Shattering Its Own Record for World’s Lowest Fertility Rate — South Korea's fertility rate fell to 0.72 expected births per woman over a lifetime from 0.78 the prior year, with births down 7.7% to 230,000 in a nation of about 50m, according to Statistics Korea. Marriages also hit a low, with about 192,500 couples wedding, down 42% from a decade earlier.
  • How a Vast Demographic Shift Will Reshape the World — In East Asia and Europe, growing numbers of retirees will depend on a shrinking working-age population, and some Asian countries are aging before becoming rich — China hit its peak working-age population at 20% of US income levels at the same point, Vietnam at 14%.
  • The Wealth of Working Nations — Because of population aging, growth in GDP per capita and GDP per working-age adult have diverged across advanced economies; from 1998 to 2019 Japan's accumulated growth per working-age adult was 31.9% versus 29.5% in the US. The authors find a standard neoclassical growth model calibrated to each country's working-age population tracks output per working-age adult in most economies.

AI Summary. Multinational companies are reporting lower revenue from China, exporting less to the country, and shipping less from it, as Chinese economic growth slowed to its weakest pace in decades and consumers spend less, particularly on foreign brands. Some firms have responded by cutting their investment in China.

Newley Purnell and Clarence Leong Wall Street Journal
Date Posted:
Is Database:
Database

Many multinational companies are exporting less from China, sending fewer products there, and seeing declines in Chinese domestic revenues. Multinationals once accounted for more than half of exports out of China, but now make up less than 1/3. @WSJ

Is China becoming less profitable for multinational companies?

Core argument: China's economic growth fell last year to its slowest rate in decades, with consumers spending less, especially on foreign brands, and the country's export machine faltering.

The Chinese market is becoming less attractive [to multinational firms]. The country’s economic growth fell to its slowest rate in decades last year; consumers there are spending less, especially on foreign brands; and its once-unstoppable export machine is faltering. The result is many multinational companies are sending less from China, exporting fewer products there, and seeing declines in their revenue from the country. The changes have prompted some firms to reduce their investment in China.

Takeaways by Macro Roundup® AI

  1. China's economic growth fell last year to its slowest rate in decades, with consumers spending less, especially on foreign brands, and the country's export machine faltering.
  2. Multinational companies are sending less from China, exporting fewer products there, and seeing declines in their China revenue as the market becomes less attractive.
  3. The changes in the Chinese market have prompted some multinational firms to reduce their investment in the country.

Related Articles:

  • China Blocks Use of Intel and AMD Chips in Government Computers — China has issued government procurement guidelines that phase out US microprocessors from Intel and AMD in state PCs and servers and sideline Microsoft's Windows and foreign database software in favor of domestic alternatives. China accounted for 27% of Intel's $54bn in sales and 15% of AMD's $23bn.
  • China Intensifies Push to ‘Delete America’ From Its Technology — A Chinese government directive requires state-owned companies in finance, energy, and other sectors to replace foreign software in their information technology systems by 2027, part of a broader push to remove U.S. technology. Hardware makers were targeted first, with Dell, IBM, and Cisco equipment gradually replaced by Chinese products.
  • Managing Economic and Financial Entanglements With China — China's surplus in manufactured goods net of commodity imports has kept growing relative to trade partners' output and relative to China's own economic output, alongside weak consumer spending and falling homebuilding. U.S. federal spending financed largely by borrowing has supported private balance sheets and demand, muting attention to that surplus.

Monday, April 1, 2024

AI Summary. The US native-born labor force fell by about 1.2m over four years, a decline driven overwhelmingly by population aging, while foreign-born labor supply grew by about 3m, leaving the overall labor force up roughly 2m. Without immigration, the total US labor force would have shrunk.

Ernie Tedeschi Briefing Book
Date Posted:
Is Database:
Database

Since the end of 2019, US GDP has grown by 8.2%, and immigration drove at least 1.6pp of that growth. The overall US workforce grew by 2mm, with immigration offsetting a 1.2mm decline in the native-born workforce. @ernietedeschi

Has immigration offset the aging-driven decline in the native-born labor force?

Core argument: The native-born US labor force fell by 1.2 million between 2019 Q4 and 2023 Q4, a decline driven overwhelmingly by aging, as the share of native-born adults aged 65 or older rose from about 21% to 23%.

Between the end of 2019 and the end of 2023, the native-born US labor force (workers who either had or actively wanted a job) fell by 1.2 million. This decline was driven overwhelmingly by the aging of the native-born population: in 2019 Q4, about 21% of the native-born civilian adult population was 65 or older; in 2023 Q4, this share had risen to 23%. In a counterfactual world without immigration, then, the overall US labor force would have shrunk over the past four years, placing a meaningful ceiling on aggregate growth. Instead, the US labor force grew by about 2 million. That is because immigrants more than filled in the labor hole dug by native-born workers. The foreign-born labor supply expanded by about 3 million, with 1 million of that from naturalized US citizens and 2 million from non-citizens, which includes immigrants with both legal and non-legal status. This immigrant contribution on its own expanded the labor force by 1.9% over the four years following 2019 Q4.

Takeaways by Macro Roundup® AI

  1. The native-born US labor force fell by 1.2 million between 2019 Q4 and 2023 Q4, a decline driven overwhelmingly by aging, as the share of native-born adults aged 65 or older rose from about 21% to 23%.
  2. Foreign-born labor supply expanded by about 3 million over the four years after 2019 Q4 — 1 million naturalized citizens and 2 million non-citizens of both legal and non-legal status — while the total US labor force grew by about 2 million.
  3. The immigrant contribution alone expanded the US labor force by 1.9% over the four years following 2019 Q4; without immigration, the overall labor force would have shrunk, placing a meaningful ceiling on aggregate growth.

Related Articles:

  • The Demographic Outlook: 2024 to 2054 — The Congressional Budget Office projects the US population will grow from 342mm in 2024 to 383mm in 2054, an average 0.4% a year — less than half the 0.9% pace of the prior five decades. Immigration accounts for about 70% of growth over the next decade and for all population growth beginning in 2040.
  • Growth in Working-Age Population Ends. That’s Not All Bad — Immigration is the only factor keeping the US working-age population from shrinking, as about 4 million baby boomers turn 65 each year and more than 820,000 Americans aged 15–64 died last year. Forecasts project immigration will hold the US 15-to-64 population roughly constant through 2100, versus projected declines of 20% across high-income countries and 62% in China.
  • Immigrants & Their Kids Were 70% of U.S. Labor Force Growth Since 1995 — U.S. labor force data show immigrants accounted for 49% of labor force growth over roughly the past three decades and their children another 21%, together about 70% of an increase of nearly 32.8m workers. Immigrants rose from about 10% to 18% of the labor force, and immigrants plus their children from 18% to 29%.

AI Summary. Nationwide chronic absenteeism — the share of students missing at least 10% of a school year — rose from 15% in 2018 to 28% in 2022, and although rates improved in 2023, they remained 75% above the pre-pandemic baseline. Chronic absenteeism is associated with lower standardized test scores, with the highest rates in lower-achieving, higher-poverty districts.

Nat Malkus American Enterprise Institute
Date Posted:
Is Database:
Database
Is Important:
Important

28% of American public school students missed at least 10% of the school year in 2022 up from 15% before the pandemic. The change was most pronounced in low-achieving school districts. @natmalkus @AEIecon

Has chronic school absenteeism remained far above pre-pandemic levels?

Core argument: Nationwide chronic absenteeism—students missing at least 10% of a school year—rose from 15% in 2018 to 28% in 2022, and although rates fell in 33 of 39 reporting states in 2023, they remained 75% above the pre-pandemic baseline.

Chronic absenteeism is associated with lower standardized test scores. Studies have shown larger associations between chronic absenteeism and achievement in later grades than in earlier grades and in math than in reading. Nationwide, chronic absenteeism—the percentage of students missing at least 10% of a school year—surged from 15% in 2018 to 28% in 2022. Falling in 33 of 39 states reporting data, chronic absenteeism rates improved in 2023 but still remained 75% higher than the pre-pandemic baseline. Chronic absenteeism increased for all district types, but rates were highest in districts with low achievement and higher poverty, affecting over one in three students. In 2022, 16% of Asian students and 24% of white students were chronically absent, compared to 36% of Hispanic students and 39% of black students. The urgent need to recover from pandemic learning loss will be severely hampered by current rates of chronic absenteeism, making it the most pressing post-pandemic problem in public schools.

Takeaways by Macro Roundup® AI

  1. Nationwide chronic absenteeism—students missing at least 10% of a school year—rose from 15% in 2018 to 28% in 2022, and although rates fell in 33 of 39 reporting states in 2023, they remained 75% above the pre-pandemic baseline.
  2. Chronic absenteeism in 2022 reached 39% of black students and 36% of Hispanic students, compared with 24% of white students and 16% of Asian students.
  3. Chronic absenteeism, which is associated with lower standardized test scores, rose across all district types but was highest in districts with low achievement and higher poverty, affecting more than one in three students there.

Related Articles:

  • What the Data Says About Pandemic School Closures, Four Years Later — Among school districts with similar poverty levels, remote learning was associated with greater declines in student achievement, and a district-level analysis found community poverty rates and the length of school closures had roughly equal effects on outcomes. For each week spent remote, students in poor districts experienced steeper math losses than peers in richer districts.
  • Parents Don’t Understand How Far Behind in School Their Kids Are — The authors calculate that the average student fell half a year behind in math and a third of a year behind in reading. Students in the poorest 10% of districts lost three-quarters of a year in math, more than double the decline in the richest districts, with reading declines half as large.
  • NAEP Long-Term Trend Assessment Results: Reading and Mathematics — U.S. national assessment scores for 13-year-olds declined 4 points in reading and 9 points in mathematics from the prior assessment, and 7 and 14 points respectively compared with a decade earlier. Mathematics declines were larger for lower-performing students (12–14 points) than for middle- and higher-performing students (6–8 points).

AI Summary. Enrollment in vocational-focused community colleges rose 16% to its highest level since tracking began, with construction-trades programs up 23% and heating, ventilation, air conditioning and vehicle repair programs up 7%. Federal data show the number of carpenters and electricians grew over the past decade even as their median ages fell.

Te-Ping Chen Wall Street Journal
Date Posted:

Since 2019 enrollment growth in vocational-focused community colleges is up 16% compared to under 1% growth in four-year colleges. The median age of American electricians and plumbers started to decline recently.

Are young workers shifting toward skilled trades and vocational training?

Core argument: Enrollment in vocational-focused community colleges rose 16% last year to the highest level since the National Student Clearinghouse began tracking the data in 2018.

The number of students enrolled in vocational-focused community colleges rose 16% last year to its highest level since the National Student Clearinghouse began tracking such data in 2018. The ranks of students studying construction trades rose 23% during that time, while those in programs covering HVAC and vehicle maintenance and repair increased 7%. The number of carpenters in the U.S. grew over the past decade, while their median age fell from 42.2 to 40.9. The same was true for electricians, whose ranks grew by 229,000 workers, even as their median age fell by 2.9 years, according to federal data. Other skilled trade occupations, such as plumbing and HVAC workers, have also trended younger.

Takeaways by Macro Roundup® AI

  1. Enrollment in vocational-focused community colleges rose 16% last year to the highest level since the National Student Clearinghouse began tracking the data in 2018.
  2. Enrollment in construction trades programs rose 23%, while programs covering HVAC and vehicle maintenance and repair grew 7%.
  3. The number of U.S. carpenters grew over the past decade as their median age fell from 42.2 to 40.9, and electricians added 229,000 workers with a median age drop of 2.9 years, according to federal data.

Related Articles:

  • America Is Trying to Electrify. There Aren’t Enough Electricians — The U.S. Bureau of Labor Statistics projects the country's more than 700,000 electricians will grow about 7% over the next decade, against a 5% average for all occupations. Electricians report being booked months out and struggling to hire, and many are raising wages and prices as federal climate incentives worth several hundred billion dollars take effect.
  • Talent Disrupted: Underemployment, College Graduates, and the Way Forward — A study classifying college graduates as underemployed when their occupation does not typically require a bachelor's degree finds that of every 100 graduates with a terminal bachelor's degree, 48 hold college-level jobs one year after graduation and 55 do at both five and 10 years. Graduates who start underemployed largely remain so.
  • China’s Shipyards Are Ready for a Protracted War. America’s Aren’t — China produced more than half of the world's commercial shipbuilding output, while Europe accounts for about 5% and the US next to nothing, with most of the remainder built in South Korea and Japan. Analysts argue that in a protracted conflict China's wartime-scale yards could quickly accelerate production, replace lost ships, and repair damaged ones.

AI Summary. The implied equity risk premium for the S&P 500 — the forward-looking extra return investors demand for holding stocks over risk-free assets — has fallen to 4.23%, its lowest level since 2008, as the market has climbed.

Aswath Damodaran Social Science Research Network
Date Posted:
Is Database:
Database

.@AswathDamodaran estimates the current equity risk premium for the S&P 500 at 4.23%, the lowest level since 2008; the expected return on stocks is 8.6%.

Has the equity risk premium fallen to its lowest level since 2008?

Core argument: The implied equity risk premium for the S&P 500 has fallen to 4.23% as the market climbs, its lowest level since 2008.

As the market climbs, the implied ERP for the S&P 500 drops to 4.23%, its lowest value since 2008. As a forward-looking price of risk, the ERP drives everything in markets. I have a review that I do on ERP, and my fifteenth annual update is now available. The paper is verbose (155 pages) and not riveting reading, but it does include everything I know about equity risk premiums and their estimation.

Takeaways by Macro Roundup® AI

  1. The implied equity risk premium for the S&P 500 has fallen to 4.23% as the market climbs, its lowest level since 2008.
  2. The equity risk premium, characterized as a forward-looking price of risk that drives everything in markets, is the subject of a fifteenth annual review running 155 pages on its determinants and estimation.

Related Articles:

  • Magnificent 7. Or is it Magnificent 2? — Microsoft and Nvidia's combined weight in the S&P 500 rose from 6.3% in 2021 to 12.3%, while Alphabet, Amazon, Apple, Meta, and Tesla's combined weight fell from 18% to 16.2% as they underperformed the index. The other 493 stocks rose 33.5% in value, from $23.9tn to $31.9tn.
  • Five Easy Pieces — The rally in the seven largest US technology stocks has been driven by earnings rather than valuation: of the group's 28% return, 21 percentage points came from sales growth, 6 from margin expansion, and 1 from multiple expansion. Market concentration has reached its highest level since 1972, and large spikes in concentration have frequently coincided with or preceded prior recessions.
  • Long-Term Shareholder Returns: Evidence From 64,000 Global Stocks — A study of about 64,000 global stocks estimates net global stock market wealth creation of $75.7tn, with returns highly concentrated: the top 159 firms (0.25%) account for half of net wealth creation and the top 1,526 (2.39%) account for all of it. Fewer than half of stocks beat one-month US Treasury bills, 42.6% of non-US and 44.8% of US stocks.

AI Summary. Market share is not a consistent indicator of firm performance, and many industries show a negative relationship between market share and profitability. Among industry-leading public firms in the dataset, 90 companies — 56% — have profit margins below their industry's median.

Daniel Rasmussen Verdad
Date Posted:
Is Database:
Database
Is Important:
Important

After examining all US and Canadian private and public firms with >$10mm in sales, @verdadcap concludes that the relationship between a firm’s market share and return on assets/equity is “virtually non-existent.”

Does higher market share mean higher profitability?

Core argument: Of the industry-leading public firms in the dataset, 90 companies—56%—have profit margins below their industry median, and it is rare for the number-one company in an industry to earn far above-market margins.

Market share is not a consistent indicator of firm performance. This finding holds true across industries, with many illustrating a negative relationship between market share and profitability. Companies earn their market share through a better product, better pricing, better service, or a litany of other firm-specific strategic decisions. It's actually quite rare to see the number-one company in an industry earning far above-market profit margins. In our dataset, 90 companies—or 56% of the industry-leading public firms—have margins below the industry median.

Takeaways by Macro Roundup® AI

  1. Of the industry-leading public firms in the dataset, 90 companies—56%—have profit margins below their industry median, and it is rare for the number-one company in an industry to earn far above-market margins.
  2. Market share is not a consistent indicator of firm performance, a finding that holds across industries, with many showing a negative relationship between market share and profitability.
  3. Companies earn market share through a better product, better pricing, better service, or other firm-specific strategic decisions.

Related Articles:

  • Rising Markups and Declining Business Dynamism: Evidence From the Industry Cross Section — Across industries, larger increases in markups are not associated with larger declines in business dynamism—measured by new business entry rates and gross job or worker flows—and the correlation may even be positive. The authors conclude that rising markups as commonly measured are not a major explanatory factor behind declining dynamism.
  • Surging Business Formation in the Pandemic: Causes and Consequences? — Shifts in consumer demand and in preferences for work and lifestyle during the pandemic created entrepreneurial opportunities, and applications to start businesses produced new firms and establishments that hired workers in large numbers. The authors argue these new businesses drew in quitting workers as founders or hires, implying economic restructuring across industry, geography, and firm size and age.
  • The Economics of Inequality in High-Wage Economies — United States Income, Wealth, Consumption, and Inequality Diana Furchtgott-Roth

AI Summary. Microsoft and OpenAI executives have drawn up plans for a US data center project, referred to as "Stargate," housing a supercomputer with millions of specialized server chips to power OpenAI's artificial intelligence. People involved say the project could cost as much as $100bn, roughly 100x some of today's biggest data centers, with Microsoft likely financing it.

Anissa Gardizy and Amir Efrati The Information
Date Posted:

Microsoft and OpenAI are planning a super-computer “Stargate” that could cost as much as $100B to provide compute for OpenAI’s artificial intelligence.

Microsoft and OpenAI are planning a super-computer “Stargate” that could cost as much as $100B to provide compute for...

How much will the largest planned AI supercomputer cost?

Core argument: Microsoft and OpenAI executives have drawn up plans for a U.S.-based data center project referred to as "Stargate" that could cost as much as $100 billion, according to people involved in the private conversations and a person who viewed Microsoft's initial cost estimates.

Executives at Microsoft and OpenAI have been drawing up plans for a data center project that would contain a supercomputer with millions of specialized server chips to power OpenAI’s artificial intelligence, according to three people who have been involved in the private conversations about the proposal. The project could cost as much as $100 billion, according to a person who spoke to OpenAI CEO Sam Altman about it and a person who has viewed some of Microsoft’s initial cost estimates. Microsoft would likely be responsible for financing the project, which would be 100 times more costly than some of today’s biggest data centers, demonstrating the enormous investment that may be needed to build computing capacity for AI in the coming years. Executives envisage the proposed U.S.-based supercomputer, which they have referred to as “Stargate,” as the biggest of a series of installations the companies are looking to build over the next six years.

Takeaways by Macro Roundup® AI

  1. Microsoft and OpenAI executives have drawn up plans for a U.S.-based data center project referred to as "Stargate" that could cost as much as $100 billion, according to people involved in the private conversations and a person who viewed Microsoft's initial cost estimates.
  2. The proposed Stargate supercomputer would contain millions of specialized server chips to power OpenAI's artificial intelligence and would be 100 times more costly than some of today's biggest data centers.
  3. Microsoft would likely be responsible for financing the project, which executives envisage as the biggest of a series of installations the two companies are looking to build over the next six years.

Related Articles:

  • AI Datacenter Energy Dilemma – Race for AI Datacenter Space — Artificial intelligence is projected to push datacenters to 4.5% of global electricity generation by 2030, with datacenter power capacity growth accelerating from a 12-15% annual rate to 25%. Global datacenter critical IT power demand is projected to rise from 49 gigawatts to 96 gigawatts, about 40 gigawatts of it for AI.
  • Electricity 2024 — More than 8,000 data centers operate globally, with about 33% in the United States, 16% in Europe, and close to 10% in China. US data center electricity use is projected to rise from roughly 200 TWh in 2022, about 4% of US electricity demand, to almost 260 TWh, or 6% of total demand.
  • Microsoft Infrastructure – AI & CPU Custom Silicon Maia 100, Athena, Cobalt 100 — Microsoft has planned annual datacenter spending of more than $50bn, which exceeds the annual spend of megaprojects such as nationwide rail networks, dams, and the Apollo moon program. The buildout is aimed at accelerating artificial general intelligence and extending generative AI from productivity applications to leisure.

AI Summary. AI adoption is argued to require continued heavy investment in semiconductor production and a significant expansion of energy infrastructure to meet rising demand for computation. This larger demand for investment could push interest rates higher; Northern Virginia's data center cluster is projected to need the equivalent of several large nuclear power plants.

Tyler Cowen Bloomberg
Date Posted:

.@tylercowen argues that demand from AI for semiconductors, computation, and energy will require capital investment that will put upward pressure on interest rates.

Will AI's demand for chips and energy push interest rates higher?

Core argument: Northern Virginia, home to major data centers, needs the equivalent of several large nuclear power plants to meet projected energy demands, a dilemma driven by AI and other factors.

Think about capex in a world of AI. The scurry to produce more high-quality semiconductor chips will continue. Those investments are not easy or cheap. But the demand for investment will not stop there. The more that AI is integrated into lives and business plans, the higher will be the demand for computation. That will induce a significant expansion of energy infrastructure. Northern Virginia, for example, is now facing a major dilemma along these lines, and not only because of AI. The region is home to major data centers, and now needs the equivalent of several large nuclear power plants to meet projected energy demands.

Takeaways by Macro Roundup® AI

  1. Northern Virginia, home to major data centers, needs the equivalent of several large nuclear power plants to meet projected energy demands, a dilemma driven by AI and other factors.
  2. Greater integration of AI into daily life and business plans raises demand for computation, which the author argues will induce a significant expansion of energy infrastructure.
  3. Production of more high-quality semiconductor chips will continue, and those investments are neither easy nor cheap.

Related Articles:

  • Amazon Bets $150 Billion on Data Centers Required for AI Boom — Amazon has committed roughly $148bn over two years to build and operate data centers worldwide, expanding hubs in northern Virginia and Oregon while entering new locations including Mississippi, Saudi Arabia, and Malaysia. The buildout targets demand for corporate services like file storage and databases, as well as the computing power for expected growth in generative artificial intelligence.
  • AI Needs So Much Power That Old Coal Plants Are Sticking Around — US utilities and grid operators have doubled their annual electricity demand forecasts for the next five years to about 1.5%, the highest rate since the 1990s, after decades of growth below 1%. Alongside data centers, the demand comes from roughly $465bn of announced semiconductor, electric-vehicle and battery factories, and old coal plants are staying online.
  • Data Centres Curbed as Pressure Grows On Electricity Grids — The International Energy Agency reports that data center energy use is growing rapidly in the US, home to a third of the world's 8,000 data centers, and in China, which has a tenth, and projects that global electricity consumption from data centers, cryptocurrencies, and artificial intelligence could double over a four-year span.

AI Summary. Newly digitized unemployment insurance claims data show U.S. states' unemployment rates have become less dispersed, with convergence in how quickly and fully unemployment recovers after recessions. The authors argue convergence in states' industrial composition helps explain this more uniform business cycle and why interstate migration became a less important adjustment mechanism.

Andrew Fieldhouse, David Munro and Christoffer Koch Brookings Papers On Economic Activity
Date Posted:
Is Database:
Database

Since the 1960s labor market fluctuations have become more uniform across states, helping to make interstate migration less important in cyclical adjustments. @A_Fieldhouse @D_Munro_Econ @ChristofferKoch

Why have state unemployment rates converged across the United States?

Core argument: Newly digitized unemployment insurance claims data, used to construct monthly state unemployment series back to January 1947, show a trend decrease in the dispersion of unemployment across U.S. states.

Using newly digitized unemployment insurance claims data we construct historical monthly unemployment series for U.S. states going back to January 1947. We document 1) a trend decrease in the dispersion of unemployment across states; 2) a marked attenuation of relative employment and relative population responses to state-specific demand shocks; and 3) a convergence across states in both the speed and degree to which unemployment recovers after recessions. Convergence in states’ industrial composition helps explain why a more uniform business cycle emerged when it did. And states’ increasingly similar experience in recessions helps explain why interstate migration became less of an important adjustment mechanism.

Takeaways by Macro Roundup® AI

  1. Newly digitized unemployment insurance claims data, used to construct monthly state unemployment series back to January 1947, show a trend decrease in the dispersion of unemployment across U.S. states.
  2. Relative employment and relative population responses to state-specific demand shocks attenuated markedly, and states converged in both the speed and degree to which unemployment recovers after recessions.
  3. Convergence in states' industrial composition helps explain the timing of a more uniform business cycle, and states' increasingly similar recession experiences help explain why interstate migration became less important as an adjustment mechanism.

Related Articles: