Edward Conard

Top Ten New York Times Bestselling Author

  • “…a must-read for serious students of economic policy.” - Glenn Hubbard, Dean, Columbia Business School, and former Chairman of the Council of Economic Advisers
  • “…a fresh argument for the productive value of inequality.” - David Autor, Professor of Economics, Massachusetts Institute of Technology
  • “…a comprehensive explanation of the modern economy.” - Julian Robertson, Founder, Tiger Management
  • “Unintended Consequences offers deep and well-argued analyses on almost every issue.” - The New York Times
  • “…challenges misconceptions that distort our economic debates.” - Arthur Brooks, President of the American Enterprise Institute
  • “Unintended Consequences is far smarter and more thought-provoking than most economics written for the general public” - Greg Mankiw, Harvard University, Former Chairman of the Council of Economic Advisors
  • “…serious thinking for serious thinkers. …a thought-provoking blueprint for growing middle- and working-class incomes.” - Mitt Romney, former Governor of Massachusetts
  • “…challenges misconceptions that distort our economic debates.” - Arthur Brooks, President of the American Enterprise Institute
  • “…a fresh argument for the productive value of inequality.” - David Autor, Professor of Economics, Massachusetts Institute of Technology
  • “A full-throated defense of economic dynamism.” - The Wall Street Journal
  • “…serious thinking for serious thinkers. …a thought-provoking blueprint for growing middle- and working-class incomes.” - Mitt Romney, former Governor of Massachusetts
  • “…a very valuable contribution.” - Larry Summers, former Secretary of the Treasury and director of the National Economic Council, president emeritus, Harvard University
Upside of Inequality Oxford Unintended Consequences

Macro Roundup


Tuesday, December 10, 2024

AI Summary. 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.

OECD Staff Organisation for Economic Co-operation and Development
Date Posted:
Is Database:
Database
Is Important:
Important

The 2023 OECD Survey of Adult Skills reveals the US has ~ 3 low-scorers for every high-scorer. Germany has nearly 3X as many high-scorers per low-scorer as the US; Scandinavia has more than 5X and Japan more than 7X.

How do adult numeracy skills vary across proficiency levels?

Core argument: Adults at Level 5 can apply problem-solving strategies to analyse, evaluate, reason, and critically reflect on complex and formal mathematical information, including dynamic representations, and can judge whether a data set supports or refutes a claim.

Adults at Level 5 can use and apply problem-solving strategies to analyse, evaluate, reason, and critically reflect on complex and formal mathematical information, including dynamic representations. They demonstrate an understanding of statistical concepts and can critically reflect on whether a data set can be used to support or refute a claim. Adults at Level 1 can interpret simple spatial representations or a scale on a map. Identify and extract information from a table or graphical representation or complete a simple whole number bar chart. Identify the largest value in an unordered list, including comparing the decimal part of the number, and interpret and perform basic arithmetic operations, including multiplication and division, with whole numbers, money, and common whole-number percentages, such as 25% and 50%. Adults performing Below Level 1 can count up to 20 objects that are displayed with varying degrees of organisation, sort events by chronological order, compare unordered lists of numbers to identify the largest number based on the whole-number component, locate data directly from a graph, and perform addition and subtraction with small whole numbers.

Takeaways by Macro Roundup® AI

  1. Adults at Level 5 can apply problem-solving strategies to analyse, evaluate, reason, and critically reflect on complex and formal mathematical information, including dynamic representations, and can judge whether a data set supports or refutes a claim.
  2. Adults at Level 1 can interpret simple spatial representations or a map scale, extract information from tables or graphs, identify the largest value in an unordered list, and perform basic arithmetic with whole numbers, money, and percentages such as 25% and 50%.
  3. Adults performing Below Level 1 can count up to 20 objects displayed with varying degrees of organisation, sort events chronologically, compare unordered lists to identify the largest whole-number value, locate data directly from a graph, and add and subtract small whole numbers.

Related Articles:

  • The Economics of Inequality in High-Wage Economies — Information technology created investment opportunities exceeding the supply of properly trained talent, while trade with low-wage economies, low-skilled immigration, trade deficits, and aging demographics eased constraints on low-skilled labor and risk-averse savings. The author argues this framework, rather than rising cronyism, explains high returns to success, slowing productivity growth, and low interest rates.
  • US PISA Results By Subpopulation — On the international PISA student assessments, American Asian and White students ranked at the top, American Hispanic students outscored Hispanic students elsewhere, and American Black students performed well. In reading, American Asian students placed first and American White students third.
  • Multidimensional Human Capital and the Wage Structure — The wage return to cognitive skills has declined, from about 10% higher hourly wages per standard deviation of test score in the 1980s and early 1990s to 4.5% in the 2000s and early 2010s, while returns to a bachelor's degree rose. Social skill-intensive occupations grew as a share of US jobs, with faster real wage growth.

AI Summary. An analysis of fund managers finds that those who outperformed the S&P 500 are clustered among managers with larger average overweight positions in the "Magnificent 7" large technology stocks. Managers that were consistently overweight each of those stocks individually showed the same pattern.

Michael Cembalest J.P. Morgan
Date Posted:

Active managers who outperformed the S&P 500 this year have been highly clustered around those who were overweight Mag 7 stocks.

Did overweighting big tech stocks drive fund manager outperformance?

Core argument: Active managers' excess returns versus the S&P 500 since January 2023 cluster around those with an average aggregate overweight to the Magnificent 7 stocks.

The first chart shows each manager’s average aggregate Mag 7 stock overweight or underweight since January 2023 on the X axis, and the manager’s excess return vs the S&P 500 on the Y axis. The outperformers are highly clustered around managers that were overweight Mag 7 stocks. We then computed the share of total months that each manager was overweight Mag 7 stocks on an individual basis; results were similar with outperformers clustered around managers that were consistently overweight each one of the Mag 7 stocks.

Takeaways by Macro Roundup® AI

  1. Active managers' excess returns versus the S&P 500 since January 2023 cluster around those with an average aggregate overweight to the Magnificent 7 stocks.
  2. Computing the share of total months each manager was overweight each Magnificent 7 stock individually produced similar results, with outperformers clustered among managers consistently overweight every one of the seven.

Related Articles:

  • Updating Our Long-Term Return Forecast For US Equities to Incorporate the Current High Level of Market Concentration — The authors estimate the S&P 500 will deliver an annualized nominal total return of 3% over the next 10 years, roughly 1% in real terms, modeling long-term returns as a function of starting valuation, market concentration, frequency of economic contractions, corporate profitability, and interest rates. The index returned 13% annualized over the past decade, above its 11% long-term average.
  • US Exceptionalism: Drivers of Equity Outperformance and What’s Needed for a Repeat — US companies need about 7% earnings-per-share growth to deliver a normal 4% equity risk premium above the risk-free rate, versus about 3% for other developed markets, with the highest expectations embedded in technology and the largest US tech firms. Chinese companies could generate a 4% risk premium over Chinese bonds with no earnings growth.
  • Which U.S. Stocks Generated the Highest Long-Term Returns? — Of the 29,078 U.S. common stocks in a database spanning nearly a century, 51.6% had negative cumulative returns, while 17 stocks returned more than 5m% — led by Altria at 265m%. Those top performers averaged annualized returns of just 13.47%, underscoring the value of time invested.

AI Summary. Recent U.S. migration has largely come through illegal channels that bypass the skills-based screening used elsewhere, and research indicates the great majority of illegal immigrants have no more than a high school education and limited English. Lower-skilled workers face higher unemployment and greater sensitivity to economic conditions.

Joseph Wang Fed Guy Blog
Date Posted:

.@FedGuy12 argues the US is “importing unemployment” given the surge in low-skilled illegal immigrants, who are particularly vulnerable to changes in business conditions.

Does low-skilled immigration leave the labor market more vulnerable to downturns?

Core argument: Research suggests the great majority of illegal immigrants in the U.S. have no more than a high school education, concentrating recent migration among lower-skilled workers, who face higher unemployment rates and greater sensitivity to economic conditions.

The lower-skilled nature of recent U.S. migrants makes the U.S. labor market particularly vulnerable to economic weakness. Unlike other countries, the recent surge in migration into the U.S. has largely been through illegal channels so the exact number is unknown. The migrants bypass the skills-based legal filters typical in other countries and tend to be lower-skilled. Research suggests that the great majority of illegal immigrants have no more than a high school education. They also predominantly originate from non-English speaking countries and so tend to have low levels of English proficiency. Lower-skilled laborers have higher unemployment rates and are also more sensitive to economic conditions. The experience abroad suggests that newer migrants have more difficulty with the labor market and that difficulty would be compounded by lower skill levels.

Takeaways by Macro Roundup® AI

  1. Research suggests the great majority of illegal immigrants in the U.S. have no more than a high school education, concentrating recent migration among lower-skilled workers, who face higher unemployment rates and greater sensitivity to economic conditions.
  2. The U.S. migration surge has come largely through illegal channels, so the exact number of arrivals is unknown and migrants bypass the skills-based legal filters typical of other countries.
  3. Recent U.S. migrants predominantly originate from non-English speaking countries and tend to have low English proficiency, and experience abroad suggests newer migrants have more difficulty in the labor market, compounded by lower skill levels.

Related Articles:

  • How Immigration Remade the U.S. Labor Force — The Congressional Budget Office estimates more than 9m people have migrated on net to the U.S. in recent years, nearly matching the prior decade's total and lifting population growth to almost 1.2% a year. Fewer than 30% are lawful permanent residents; without immigration, growth would be 0.2%.
  • Immigration Surge Has Slowed: Updated Estimates of Net International Migration — The authors estimate that net international migration to the US was close to 3.5m in one year — about 200,000 above the Congressional Budget Office's estimate, driven largely by higher undocumented entries — and project roughly 2.9m the following year, about 400,000 below the Congressional Budget Office's projection.
  • 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.

AI Summary. Researchers report superconducting quantum processors running surface code memories below the error-correction threshold, with the logical error rate suppressed 2.14x per two-step increase in code distance and a 101-qubit code reaching 0.143% error per cycle. The logical memory outlived its best physical qubit by 2.4x.

Rajeev Acharya, Dmitry Abanin, Laleh Aghababaie-Beni, et al. Nature
Date Posted:

Google is releasing Willow, a quantum computing chip that can solve in less than 5 minutes a computation that would take the current leading supercomputer over 10 septillion years (10^25) “a number that vastly exceeds the age of the Universe.”

Can quantum error correction suppress logical errors as code size grows?

Core argument: The Willow superconducting processor runs a 101-qubit distance-7 surface code memory with 0.143% ± 0.003% error per error-correction cycle, with logical error suppressed by a factor of Λ=2.14±0.02 when the code distance increases by two.

We present two below-threshold surface code memories on our newest generation of superconducting processors, Willow: a distance-7 code, and a distance-5 code integrated with a real-time decoder. The logical error rate of our larger quantum memory is suppressed by a factor of Λ=2.14±0.02 when increasing the code distance by two, culminating in a 101-qubit distance-7 code with 0.143% ± 0.003% error per cycle of error correction. This logical memory is also beyond break-even, exceeding its best physical qubit’s lifetime by a factor of 2.4±0.3. Our system maintains below-threshold performance when decoding in real time, achieving an average decoder latency of 63 μs at distance-5 up to a million cycles, with a cycle time of 1.1 μs. We also run repetition codes up to distance-29 and find that logical performance is limited by rare correlated error events occurring approximately once every hour, or 3×109 cycles. Our results present device performance that, if scaled, could realize the operational requirements of large-scale fault-tolerant quantum algorithms.

Takeaways by Macro Roundup® AI

  1. The Willow superconducting processor runs a 101-qubit distance-7 surface code memory with 0.143% ± 0.003% error per error-correction cycle, with logical error suppressed by a factor of Λ=2.14±0.02 when the code distance increases by two.
  2. The distance-7 logical memory operates beyond break-even, exceeding the lifetime of its best physical qubit by a factor of 2.4±0.3.
  3. Repetition codes run up to distance-29 show logical performance limited by rare correlated error events occurring approximately once every hour, or 3×10⁹ cycles, while real-time decoding at distance-5 averages 63 μs latency against a 1.1 μs cycle time.

Related Articles:

  • Meet Willow, Our State-of-the-Art Quantum Chip and — Google reports that its Willow quantum chip completed a random circuit sampling benchmark computation in under five minutes that one of today's fastest supercomputers would take an estimated 10^25 years to perform. The benchmark is described as the classically hardest test currently runnable on a quantum computer.
  •  Decoding For Quantum Processors — Researchers developed a recurrent, transformer-based neural network that decodes the surface code, the leading quantum error-correction code, and report it outperforms other state-of-the-art decoders on data from Google's Sycamore quantum processor and on simulated data with realistic noise. The authors argue machine learning can surpass human-designed decoding algorithms by learning directly from data.
  • Quantum Computing Advance Begins New Era, IBM Says — IBM researchers report that a 127-qubit quantum processor simulated the magnetic behavior of 127 atom-scale bar magnets, a problem too complex for an exact answer on the largest supercomputers. Because quantum noise made each sub-millisecond calculation unreliable, the computation was repeated 600,000 times to converge on an answer.

AI Summary. Google reports that its Willow quantum chip completed a random circuit sampling benchmark computation in under five minutes that one of today's fastest supercomputers would take an estimated 10^25 years to perform. The benchmark is described as the classically hardest test currently runnable on a quantum computer.

Hartmut Neven Google
Date Posted:

Google’s Hartmut Neven notes that the performance of Willow, their new quantum computing chip, “lends credence to the notion that quantum computation occurs in many parallel universes, in line with the idea that we live in a multiverse.”

Can a quantum chip outperform supercomputers on the hardest known benchmark?

Core argument: Google's Willow quantum chip completed a random circuit sampling computation in under five minutes that, according to Google, would take one of today's fastest supercomputers 10^25 years, vastly exceeding the age of the universe.

As a measure of Willow’s performance, we used the random circuit sampling (RCS) benchmark. Pioneered by our team and now widely used as a standard in the field, RCS is the classically hardest benchmark that can be done on a quantum computer today. We’ve consistently used this benchmark to assess progress from one generation of chip to the next — we reported Sycamore results in October 2019 and again recently in October 2024. Willow’s performance on this benchmark is astonishing: It performed a computation in under five minutes that would take one of today’s fastest supercomputers 1025 or 10 septillion years. This mind-boggling number exceeds known timescales in physics and vastly exceeds the age of the universe. It lends credence to the notion that quantum computation occurs in many parallel universes, in line with the idea that we live in a multiverse, a prediction first made by David Deutsch.

Takeaways by Macro Roundup® AI

  1. Google's Willow quantum chip completed a random circuit sampling computation in under five minutes that, according to Google, would take one of today's fastest supercomputers 10^25 years, vastly exceeding the age of the universe.
  2. Random circuit sampling, pioneered by Google's team and now a standard in the field, is the classically hardest benchmark that can be run on a quantum computer today, with Sycamore results reported in October 2019 and October 2024.
  3. Google argues Willow's benchmark result lends credence to the notion that quantum computation occurs in many parallel universes, in line with the multiverse prediction first made by David Deutsch.

Related Articles:

  • Quantum Error Correction Below the Surface Code Threshold — Researchers report superconducting quantum processors running surface code memories below the error-correction threshold, with the logical error rate suppressed 2.14x per two-step increase in code distance and a 101-qubit code reaching 0.143% error per cycle. The logical memory outlived its best physical qubit by 2.4x.
  • Learning High-Accuracy Error Decoding For Quantum Processors — Researchers developed a recurrent, transformer-based neural network that decodes the surface code, the leading quantum error-correction code, and report it outperforms other state-of-the-art decoders on data from Google's Sycamore quantum processor and on simulated data with realistic noise. The authors argue machine learning can surpass human-designed decoding algorithms by learning directly from data.
  • Quantum Computing Advance Begins New Era, IBM Says — IBM researchers report that a 127-qubit quantum processor simulated the magnetic behavior of 127 atom-scale bar magnets, a problem too complex for an exact answer on the largest supercomputers. Because quantum noise made each sub-millisecond calculation unreliable, the computation was repeated 600,000 times to converge on an answer.

AI Summary. Using matched employer-employee data on Italian firms, the authors estimate that family firms pay substantially lower wages, with about half the gap (roughly 8%) explained by workers' fixed characteristics and the remainder unexplained by sorting, rent-sharing, productivity, industry, or location. Family firms also have fewer management positions, promote less, and offer lower returns to promotion.

Edoardo Di Porto, Marco Pagano, Vincenzo Pezone, et al. National Bureau of Economic Research
Date Posted:

A 14-year study of 20mm Italian workers finds that family-owned firms have ~ 16% lower wages than non-family-owned firms, not due to differential productivity or amenities. Fewer promotions in family firms suggest efforts to maintain family control.

Do family firms pay lower wages and offer fewer promotions?

Core argument: Using matched employer-employee data on Italian firms, the authors find family firms pay substantially lower wages, with about half the gap — roughly 8 log-points (~8%) — attributable to workers' time-invariant characteristics.

Based on a much larger sample of Italian firms than previous literature, we confirm the finding of a significant salary gap, family firms paying substantially lower wages. Our comprehensivematched employer-employee data permit exploration of the possible reasons. We find that about half of the gap, corresponding to 8 log-points [~ 8%] is due to workers’ time-invariant characteristics, but the rest cannot be explained by sorting, differences in rent-sharing, productivity, industry, and location. The difference is greater in high-productivity firms and is driven mainly by differences in compensation for relatively high-paid employees, i.e., those in the top percentilesin each firm’s earnings distribution. In addition, controlling for observable worker and firm characteristics, family firms tend to have fewer top and middle management positions, to promote less, and to offer lower returns to promotion. We [also] show that productivity is undercut more severely following the death of the CEO in family firms suggest[ing] the controlling family may be reluctant to share organizational and technical know-how with non-relatives.

Takeaways by Macro Roundup® AI

  1. Using matched employer-employee data on Italian firms, the authors find family firms pay substantially lower wages, with about half the gap — roughly 8 log-points (~8%) — attributable to workers' time-invariant characteristics.
  2. The remaining family-firm wage gap is not explained by sorting, rent-sharing differences, productivity, industry, or location, and is larger in high-productivity firms and driven mainly by employees in the top percentiles of each firm's earnings distribution.
  3. Controlling for observable worker and firm characteristics, family firms have fewer top and middle management positions, promote less, and offer lower returns to promotion, and their productivity falls more sharply after a CEO's death, suggesting reluctance to share know-how with non-relatives.

Related Articles:

  • Capitalists in the Twenty-First Century — Most income at the top of the US distribution is non-wage income, primarily profits from closely-held, mid-market firms in skill-intensive industries owned by working-age individuals. The authors find private business profit falls by three-quarters after an owner retires or dies prematurely, indicating most top earners are "working rich" deriving income from human capital.
  • Public Markets Are a Small Part of the Overall Economy — Publicly traded companies represent a small share of US economic activity: 87% of firms with revenues above $100m are private, privately owned firms account for nearly 80% of job openings, and S&P 500 companies provide 18% of US employment and 15% of capital spending.
  • The Market for CEOs: Evidence from Private Equity — In a study of larger U.S. companies bought by private equity, more than 70% hired new chief executives, and over 75% of those were external hires, unlike public companies where internal promotions dominate. The authors estimate these executives' pay exceeds that at similar-sized public companies, suggesting an active market for chief executives.

AI Summary. Tariff policy succeeds, the author argues, if it raises domestic production enough to pull consumption up with it, leaving American consumers better off even as consumption's share of domestic production falls and the trade deficit narrows. It fails if production does not rise, in which case consumption's share falls mainly because consumption itself declines.

Michael Pettis Financial Times
Date Posted:

.@michaelxpettis argues that a tariff policy is successful “if it raises domestic production by enough to pull consumption up with it.”

Do tariffs make consumers better off only by raising domestic production?

Core argument: Tariff policy succeeds, in this framing, only if it raises domestic production enough to pull consumption up with it, leaving American consumers better off even as consumption's share of domestic production declines.

Tariff policy is “successful”, in other words, if it raises domestic production by enough to pull consumption up with it — ie, if it causes Americans to consume more by producing even more. In that case American consumers are clearly better off, even as the share they consume of total domestic production declines. Of course, as production rises faster than consumption, this usually means that the trade deficit is declining. On the other hand, tariff policy is a “failure” if it doesn’t cause a rise in domestic production, in which case tariffs reduce the consumption share of GDP mainly by causing consumption to fall. This would clearly make American consumers worse off.

Takeaways by Macro Roundup® AI

  1. Tariff policy succeeds, in this framing, only if it raises domestic production enough to pull consumption up with it, leaving American consumers better off even as consumption's share of domestic production declines.
  2. When production rises faster than consumption under successful tariff policy, the trade deficit usually declines.
  3. Tariff policy fails if domestic production does not rise, in which case consumption's share of GDP falls mainly because consumption itself falls, leaving American consumers worse off.

Related Articles:

  • Fiscal, Macroeconomic, and Price Estimates of Tariffs Under Both Non-Retaliation and Retaliation Scenarios — The authors estimate the modeled tariff proposals would raise $1.2tn–$4.4tn over 10 years, or 0.3–1.2% of average GDP, with foreign retaliation cutting that revenue by 12–26%. They project consumer prices initially rise 1.2–5.1%, equal to $1,900–$7,600 of lost disposable income per household, with real GDP contracting 0.5–1.4% in the medium term.
  • Can Trump Reduce the Trade Deficit? — A world of widespread tariffs would also shrink international capital flows, and because the US trade deficit is the counterpart of large capital inflows, reduced capital mobility would narrow that deficit. The author argues this would occur largely by reducing overall investment in the US economy.
  • Trade Intervention for Freer Trade — The authors argue the United States should counter surplus countries' trade policies either by restricting their goods exports through tariffs or by restricting their capital inflows through taxes, volume limits, reserve requirements, transaction approvals, or investor eligibility rules. They contend capital controls are a more targeted alternative to uniform tariffs, which impose equal costs on more and less mercantilist countries.

Wednesday, December 11, 2024

AI Summary. The authors find persistent performance differences among angel investors are consistent with better-performing angels having broader access to investment opportunities: top angels are less likely to invest in tight-knit clusters and are more likely to be connected to startup board members and founders before their first investment.

Johan Karlsen, Katja Kisseleva, Aksel Mjøs and David Robinson National Bureau of Economic Research
Date Posted:
Is Database:
Database

Angel investments in Norway from 2004-18 had highly skewed results: ⅓ were a total loss, the top 1% returned 50X invested capital, and the average investment returned 2X. Top angels tended to be connected to board members and founders before investing.

Do better-performing angel investors owe their success to broader deal access?

Core argument: Persistent performance differences among angel investors stem partly from better access to investment opportunities, such that top angels would outperform even if they allocated their investments randomly, the authors find evidence consistent with this explanation.

Persistent individual differences are critical for understanding this market. One explanation for these persistent differences is that some angel investors have better access to investments than others, such that, even if they were to allocate their investments randomly, they would do better. We find evidence consistent with this explanation. Top angels are less likely than worse-performing angels to invest in tight-knit clusters. This runs counter to the idea that a small group of angels has privileged access to high-quality deals from which others are potentially excluded. We find that angels who later go on to become top angels in our sample are likelier to be connected to board members and founders of startups before they make their first investment. The greater breadth of top angels’ co-investor base and their startup network allow us to deduce that some of the unexplained component of their performance derives from their larger pools from which to draw investment opportunities.

Takeaways by Macro Roundup® AI

  1. Persistent performance differences among angel investors stem partly from better access to investment opportunities, such that top angels would outperform even if they allocated their investments randomly, the authors find evidence consistent with this explanation.
  2. Top angel investors are less likely than worse-performing angels to invest in tight-knit clusters, which runs counter to the idea that a small group of angels holds privileged access to high-quality deals excluding others.
  3. Angels who later become top performers are likelier to be connected to startup board members and founders before making their first investment, and their broader co-investor and startup networks provide larger pools of investment opportunities.

Related Articles:

  • Risk-Adjusting the Returns to Venture Capital — The authors estimate that venture capital investments in start-ups earn substantial positive risk-adjusted returns while venture capital funds' risk-adjusted returns are near zero. The systematic part of these payoffs resembles the negatively skewed returns from selling index put options; including those in the pricing model turns fund returns negative.
  • The Power Law Review: Chasing Unicorns — The median venture-capital fund underperformed the S&P 500 over roughly four decades, while the top 5% of funds returned nearly 3x the index, reflecting how a small number of deals account for most returns. Venture-backed companies made up nearly half of U.S. nonfinancial initial public offerings over a comparable period.
  • 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. More than 20% of new venture investments on the Carta platform were down rounds, in which startups raise capital at reduced valuations, one of the highest quarterly shares since at least 2019. Venture-backed valuations have undergone a reset after reaching all-time highs, and the average interval between priced rounds has lengthened.

Ashley Neville and Kevin Dowd Carta
Date Posted:

.@cartainc finds that venture valuations have undergone a reset in the last two years: between Q2 2022 and Q3 2024, 18% of new venture investments were “down rounds,” up from 10% between 2019 and Q1 2022.

Are startup valuations still resetting downward in private markets?

Core argument: More than 20% of new venture investments on Carta in Q3 2024 were down rounds, the fourth-highest quarterly total dating back to at least the start of 2019.

More than 20% of all new venture investments on Carta in Q3 were down rounds, the fourth-highest quarterly total dating back to at least the start of 2019. Startups are still grappling with the realities of the widespread valuation reset that began in mid-2022. During 2021 and early 2022, venture-backed valuations soared to all-time highs after nearly a decade of steady growth. It was rare to see any startup raise new capital at a reduced valuation. But as valuations have undergone a reset in the past two-plus years, some companies have found it difficult to raise capital at the lofty valuations they once attained. Given how the average time interval between priced rounds has also increased in the past two years, many of the companies that raised capital in Q3 were doing so for the first time since valuations began to decline.

Takeaways by Macro Roundup® AI

  1. More than 20% of new venture investments on Carta in Q3 2024 were down rounds, the fourth-highest quarterly total dating back to at least the start of 2019.
  2. The average interval between priced rounds has lengthened over the past two years, so many companies raising capital in Q3 2024 were doing so for the first time since valuations began falling, according to Carta.

Related Articles:

  • AI Investments Are Booming, but Venture-Firm Profits Are at a Historic Low — U.S. venture firms returned $26bn worth of shares to their investors in a year, the lowest amount since 2011, according to data provider PitchBook, and invested $60bn more than they collected — the largest such deficit in 26 years of PitchBook data. Investment remains high while acquisitions and initial public offerings are few.
  • Start-Up Failures Rise 60% As Founders Face Hangover From Boom Years — Data from private-company services provider Carta show start-up shutdowns rising sharply, with 254 venture-backed clients going bust in a single quarter and the failure rate more than 7x the level when tracking began. Carta also reports a steep drop in the share of companies able to raise new funding within two years of their last round.
  • Monetary Policy and Innovation — The authors estimate that a 100bps monetary tightening reduces investment in intellectual property by about 1%, research and development spending at public firms by about 3%, venture capital investment by up to 25%, and patenting in important technologies by up to 9%.

AI Summary. Foreign-held assets in Mexico, including foreign direct investment, have declined relative to Mexico's GDP, with new foreign direct investment falling to a 10-year low despite attention to nearshoring. Reinvested earnings now make up most inflows, and manufacturing absorbed nearly half of total foreign direct investment.

Enrique Martínez García, Manuel Sánchez and Luis Torres Federal Reserve Bank of Dallas
Date Posted:

A @DallasFed analysis finds that foreign investment in Mexico has declined as a % of GDP since 2022. The impact of “nearshoring” on Mexico’s economy has been limited, as “large-scale foreign capital relocation has yet to materialize.”

Has nearshoring attention translated into new foreign investment in Mexico?

Core argument: Mexico's new FDI has fallen to a 10-year low since 2022, and international investment position data show foreign-held assets including FDI relative to GDP have declined over the same period.

Despite the increased attention to nearshoring, foreign-held assets (including FDI) relative to Mexico’s GDP have declined since 2022, international investment position data show. While new FDI surged immediately following the pandemic-era recession, new FDI since 2022 has declined to a 10-year low. Foreign capital inflows from company parent accounts and non-FDI portfolio investments have also been limited. Reinvested earnings now account for most FDI inflows, reflecting a shift toward building additional capacity from own resources. The manufacturing sector absorbed nearly half of total FDI between early 2022 and mid-2024, with key contributions from the transportation equipment, food processing and metals industries. The U.S. remains Mexico’s largest source of FDI.

Takeaways by Macro Roundup® AI

  1. Mexico's new FDI has fallen to a 10-year low since 2022, and international investment position data show foreign-held assets including FDI relative to GDP have declined over the same period.
  2. Reinvested earnings account for most of Mexico's FDI inflows, reflecting a shift toward building additional capacity from own resources, while inflows from company parent accounts and non-FDI portfolio investments have been limited.
  3. Manufacturing absorbed nearly half of Mexico's total FDI between early 2022 and mid-2024, with key contributions from transportation equipment, food processing and metals, and the U.S. remains Mexico's largest source of FDI.

Related Articles:

  • Mexico Gets Cold Feet Over New Chinese EV Plant After Trump Win — Chinese electric-vehicle maker BYD is finalizing plans for a factory in Mexico, but Mexican officials worry the plant would signal to the incoming US administration that Mexico is a backdoor for Chinese companies selling into the United States. Mexico shipped nearly 2.6m vehicles to the US last year.
  • Mexico Wants to Curb Chinese Imports With Help From U.S. Companies — Mexico is asking major foreign manufacturers operating in the country—including U.S. carmakers, semiconductor firms, and aerospace and electronics companies—to identify Chinese products and parts that could be made locally, aiming to reduce reliance on imports from China, Malaysia, Vietnam, and Taiwan. China accounts for 20% of Mexico's total imports, according to Mexican government data.
  • Mexico’s Investment Boom — Mexico's export boom is largely genuine rather than a channel for tariff avoidance, since much of the growth comes from industries like motor vehicles where China has little presence in the North American market. Mexican manufacturing output has risen 5% since 2018, with computer output up nearly 25% and appliance and electrical equipment output up 15%.

AI Summary. South Korea is the leading foreign investor in new US greenfield facilities, led by auto and electronics manufacturers including Hyundai, Kia, LG and Samsung, with $21bn in commitments in the most recent year. The investment has landed in smaller cities such as Kokomo, Indiana, Savannah, Georgia and Auburn, Alabama, rather than the coastal metropolitan areas where Korean-Americans have traditionally settled.

Amanda Chu Financial Times
Date Posted:

Korean firms made $21B of commitments to greenfield projects in the United States last year making Korea the leading foreign investor in greenfield facilities in the United States. Alongside the investment small ‘Koreatowns’ have been springing up.

Is South Korean manufacturing investment creating Koreatowns in smaller American cities?

Core argument: South Korea has become the leading foreign investor in US greenfield facilities, with commitments totalling $21bn last year, the highest level in more than a decade.

Korea has become the leading foreign investor in greenfield facilities in the US, led by its auto and electronics manufacturers such as Hyundai, Kia, LG and Samsung, with commitments totalling $21bn last year, the highest level in more than a decade. But the towns that have benefited are far from the large coastal metropolitan centres where Korean-Americans have traditionally immigrated. In cities like Kokomo, and Savannah in Georgia, and Auburn in Alabama, Korean-Americans have brought their customs and cultures to unexpected corners of the country.

Takeaways by Macro Roundup® AI

  1. South Korea has become the leading foreign investor in US greenfield facilities, with commitments totalling $21bn last year, the highest level in more than a decade.
  2. Korean auto and electronics manufacturers including Hyundai, Kia, LG and Samsung have led the greenfield investment push into the United States.
  3. Kokomo, Savannah in Georgia and Auburn in Alabama lie far from the large coastal metropolitan centres where Korean-Americans have traditionally immigrated, yet Korean-Americans have brought their customs and cultures to these towns.

Related Articles:

  • Republican Districts Dominate US Clean Technology Investment Boom — A Financial Times analysis found that more than 80% of pledged investment in large-scale clean energy and semiconductor manufacturing is destined for Republican congressional districts. The analysis attributes this partly to those districts' available land and cheap labor, and to state-level tax breaks and subsidies in states such as Georgia and Ohio.
  • A US Industrial Renaissance Has Started — US legislation supporting semiconductors, clean energy, and infrastructure has triggered an industrial renaissance in artificial intelligence and energy, and US manufacturing capacity is now growing after declining for decades.
  • Business Spending Is Lining Up to Reinforce Growth — Around 70% of sectors are posting positive year-over-year profit growth, with big tech showing the largest acceleration in demand and profits but strength broadening across the economy. Companies are responding with rising capital spending growth and more optimistic investment plans, as firms typically add capacity with a slight lag after real demand picks up.

AI Summary. US math scores on the international Trends in Mathematics and Science Study fell 18 points for fourth graders and 27 points for eighth graders, on a scale where about 30 points equates to a year of learning. Declines were concentrated among the lowest-performing students, with fourth-grade scores above the 75th percentile barely moving.

Nat Malkus American Enterprise Institute
Date Posted:

Young Americans’ test scores have dropped significantly since 2011, with the lowest-performing students showing the steepest declines. @natmalkus notes that this trend pre-dates the pandemic.

Are falling US math scores concentrated among the lowest-performing students?

Core argument: US math scores on the 2023 Trends in Mathematics and Science Study fell 18 points for fourth graders and 27 points for eighth graders versus 2019, on a scale where 30 points roughly equates to a year of learning.

2023 scores from the Trends in Mathematics and Science Study (TIMSS)—an international assessment measuring fourth and eighth graders in math and science—offers a fresh look at the academic performance of US and international students. The results are grim. The pandemic harmed student achievement. Between 2019 and 2023, US math scores cratered, dropping by 18pp for fourth graders and a whopping 27pp for eighth graders—huge declines on a scale where 30 points roughly equates to a year of learning. In science, declines weren’t statistically significant but were in the same direction. The lowest performing students were hit hardest. Looking beneath the averages reveals how students across percentiles fared in both grades and subjects, and the patterns—though not all statistically significant—are consistent and compelling. In fourth-grade math, the scores of students below the 50th percentile dropped precipitously while the 75th and 90th percentile hardly moved at all.

Takeaways by Macro Roundup® AI

  1. US math scores on the 2023 Trends in Mathematics and Science Study fell 18 points for fourth graders and 27 points for eighth graders versus 2019, on a scale where 30 points roughly equates to a year of learning.
  2. US science declines on TIMSS between 2019 and 2023 were not statistically significant but moved in the same direction as the math drops.
  3. In fourth-grade math on TIMSS, scores for students below the 50th percentile dropped precipitously while the 75th and 90th percentiles hardly moved, part of a consistent pattern across grades and subjects that is not always statistically significant.

Related Articles:

  • Do Adults Have the Skills They Need to Thrive in a Changing World? — An international assessment of adult skills describes numeracy proficiency levels ranging from counting objects and adding small whole numbers at the lowest level, to extracting information from tables and graphs and performing basic arithmetic with money and percentages at Level 1, up to analysing complex mathematical information and critically evaluating whether data support a claim at Level 5.
  • The Latest Chronic Absenteeism Numbers — Chronic absenteeism—students missing 10% or more of the school year—rose from 15% of K–12 students before the pandemic to 28%, then eased to 26% in the first full post-pandemic school year. Data from eight states show continued declines, including New Mexico's drop from 40% to 33%, but rates remain well above pre-pandemic levels.
  • Students Are Making a ‘Surprising’ Rebound From Pandemic Closures. But Some May Never Catch Up — An analysis of standardized test data led by Stanford and Harvard researchers finds elementary and middle-school students have regained about a third of pandemic learning losses in math and a quarter in reading. Because poorer districts lost more ground initially, the typical poor district remains nearly half a grade level behind, versus about a fifth for rich districts.