Edward Conard

Top Ten New York Times Bestselling Author

  • “…serious thinking for serious thinkers. …a thought-provoking blueprint for growing middle- and working-class incomes.” - Mitt Romney, former Governor of Massachusetts
  • “…reminds us that inequality sends a signal of what society lacks most, in America’s case, entrepreneurship and risk taking.” - Lawrence Lindsey, CEO, The Lindsey Group, former Director of the National Economic Council
  • “Unintended Consequences offers deep and well-argued analyses on almost every issue.” - The New York Times
  • “…a fresh argument for the productive value of inequality.” - David Autor, Professor of Economics, Massachusetts Institute of Technology
  • “…a must-read for serious students of economic policy.” - Glenn Hubbard, Dean, Columbia Business School, and former Chairman of the Council of Economic Advisers
  • “Unintended Consequences is full of substance, it is one of the must-read books of the year, and once I finish it I will be giving it a second read through right away.” - Tyler Cowen, Professor, George Mason University
  • “…a comprehensive explanation of the modern economy.” - Julian Robertson, Founder, Tiger Management
  • “Unintended Consequences should be read by anyone who takes for granted the superiority of progressive taxation and has not thought carefully about the trade-offs involved.” - The New Republic
  • “There are an amazing number of good ideas and interesting points made in Unintended Consequences. The thinking underlying it, and the obvious depth of understanding of the author, are very impressive.” - Steven Levitt, coauthor of Freakonomics; 2004 John Bates Clark Medal
  • “Unintended Consequences is far smarter and more thought-provoking than most economics written for the general public” - Greg Mankiw, Harvard University, Former Chairman of the Council of Economic Advisors
  • “Unintended Consequences represents the most cogent and persuasive analysis of the Financial Crisis to date.” - Andrei Shleifer, 1999 John Bates Clark Medal Winner
  • “…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


Thursday, May 9, 2024

AI Summary. Studying changes in six U.S. state corporate income taxes, the authors find that higher corporate taxes on average depress leverage at small private firms, whose borrowing rises after tax cuts, while large private firms' leverage does not respond. Bank assessments of loan default rates fall 20–30 basis points following tax cuts.

Ivan Ivanov, Luke Pettit, and Toni Whited National Bureau of Economic Research
Date Posted:
Is Database:
Database

.@toniwhited and @ivan_t_ivanov find that state tax cuts “make capital more productive and spur borrowing” by small private firms, as investment opportunities and lower credit spreads outweigh the reduced tax subsidy to borrowing.

Do higher corporate taxes reduce borrowing at small private firms?

Core argument: Studying changes in 6 state corporate income taxes since the 1990s across private and public companies, the authors find corporate taxes on average depress leverage at small private firms, while large private firms' leverage does not respond to tax legislation.

[Miller-Modigliani theory says] the benefit from the interest tax deduction is larger than countervailing costs of debt related to financial distress, and prior empirical work finds a positive relation between debt and taxes. Using comprehensive samples of both private and public companies, we study the evolution of corporate borrowing around changes in 6 state corporate income taxes since the 1990s. For small private firms corporate taxes, on average, depress leverage. The leverage of large private firms does not change in response to tax legislation. Small firms' leverage first responds to the tax cuts in the enactment year with no significant reaction before enactment. The event study coefficients on or after the enactment dates [Figure a, panels A and B] are precisely estimated, ranging from approximately 4% in the enactment year to about 2% four years later. We find that bank assessments of loan default rates fall by 20-30bps after tax cuts. This last result supports the notion that default risk [is] also tax-sensitive.

Takeaways by Macro Roundup® AI

  1. Studying changes in 6 state corporate income taxes since the 1990s across private and public companies, the authors find corporate taxes on average depress leverage at small private firms, while large private firms' leverage does not respond to tax legislation.
  2. Small private firms' leverage first responds in the enactment year of tax cuts, with event-study coefficients ranging from roughly 4% in the enactment year to about 2% four years later and no significant reaction beforehand.
  3. Bank assessments of loan default rates fall by 20–30bps after tax cuts, which the authors take as support for the notion that default risk is tax-sensitive.

Related Articles:

  • Tax Policy and Investment in a Global Economy — Using administrative tax data, the authors estimate that firms experiencing the mean corporate tax change under the 2017 US tax law increased investment by 20% relative to firms with no change, with long-run domestic and total capital up 7% and 13%. They find higher depreciation deductions largely offset added revenue from capital accumulation.
  • Who Gains from Corporate Tax Cuts? — The authors estimate that corporate tax cuts boost overall investment and employment, with goods-producing firms such as manufacturers raising capital expenditure and wage bills without changing dividends, and a 1% cut in the marginal tax rate lifting their capital spending by as much as 8%. Service-sector firms instead direct most of the windfall to dividends.
  • End of an Era: The Coming Long-Run Slowdown in Corporate Profit Growth and Stock Returns — The authors estimate that declining interest and corporate tax rates accounted for over 40% of real corporate profit growth from 1989 to 2019, while falling risk-free rates explain the entire expansion in price-to-earnings multiples. Real earnings before interest and taxes grew slightly more slowly than in 1962–1989, at 2.2% versus 2.4% per year.

AI Summary. China's outward investment flows show rising outflows to countries across Asia and, on a smaller scale, Europe, consistent with shifting production to third countries whose exports are not automatically subject to U.S. tariffs. Gaps between China's two official foreign direct investment series may reflect losses at foreign-owned subsidiaries rather than asset sales by foreigners.

Robin Brooks Brookings Institution
Date Posted:

.@robin_j_brooks argues the reduction in net Chinese FDI since 2021 was driven primarily by an increase in outbound Chinese FDI in countries not subject to US tariffs, not by repatriation of capital by foreign investors in China.

Do China's outward investment flows show production shifting to other countries?

Core argument: Reinvested earnings accounted for 80% of the gap between China's broader SAFE foreign direct investment data and the Ministry of Commerce's narrower greenfield proxy in the few years for which such data are available.

China is likely diversifying its manufacturing base, shifting production to third countries where exports will not automatically be subject to U.S. tariffs. This is apparent in China’s outward investment flows, which show a rise in outflows to countries across Asia and—on a smaller scale—to countries in Europe. This shift is consistent with China financially decoupling from the U.S. and diversifying manufacturing supply chains at the same time. Nick Lardy has flagged that the difference between non-resident FDI flows from SAFE versus FDI data from the Ministry of Commerce implies large asset sales in 2023, as MofCom data proxy for “greenfield” FDI while SAFE data are broader. However, for the few years in which data on reinvested earnings are available they accounted for 80% of the gap between SAFE and MofCom data. As reinvested earnings are not an actual capital flow—they are booked as an offsetting dividend payment in the current account and an FDI inflow in the financial account—it is possible that a big part of the 2023 outflow reflects losses at foreign-owned subsidiaries, consistent with lingering effects from China’s COVID lockdowns, and thus has nothing to do with asset sales by foreigners.

Takeaways by Macro Roundup® AI

  1. Reinvested earnings accounted for 80% of the gap between China's broader SAFE foreign direct investment data and the Ministry of Commerce's narrower greenfield proxy in the few years for which such data are available.
  2. China's outward investment flows show rising outflows to countries across Asia and, on a smaller scale, to Europe, consistent with shifting production to third countries whose exports are not automatically subject to U.S. tariffs.
  3. Nick Lardy has flagged that the SAFE–Ministry of Commerce divergence implies large asset sales in 2023, though a big part of that outflow may instead reflect losses at foreign-owned subsidiaries, consistent with lingering effects of China's COVID lockdowns.

Related Articles:

  • Foreign Direct Investment is Exiting China, New Data Show — Two official Chinese data series—one measuring gross foreign direct investment inflows and one measuring inflows minus outflows—both show foreign direct investment exiting China. The author attributes most of the $113bn gap between the two measures to net sales of direct investment assets by foreign investors.
  • The Rise & Fall of Foreign Direct Investment in China — Net foreign direct investment into China has been negative for four consecutive quarters, as foreign investors shifted from reinvesting earnings from their Chinese operations to repatriating them. Whether this reflects cash management or signals a future slowdown in foreign direct investment is unclear, but the repatriation of profits is negative for China's currency.
  • China’s Unremitted Export Earnings — A substantial share of China's net export proceeds is never repatriated and cannot be accounted for; the authors attribute this to offshore holdings by private Chinese firms, overseas accumulation by state-owned entities, and reduced reinvestment by foreign exporters. They argue state-linked funds could be repatriated in a crisis, while lower foreign reinvestment is more structural.

AI Summary. An analysis of over one million television transcripts finds that coverage of gasoline prices shows a negative news bias, ramping up sharply once the nominal price reaches $3.50 a gallon, most steeply at Fox News. Because that threshold is fixed in nominal terms, the inflation-adjusted trigger has fallen by one-third.

Ryan Cummings, Giacomo Fraccaroli, and Neale Mahoney 
Date Posted:

.@weakinstrument @nealemahoney @giacomofcr find that between 2004-23, US television ramps up coverage of gas prices when they hit $3.50/gallon. Currently, the mean price of a gallon of gas is $3.60 nationally.

Does television coverage of gasoline prices show a negative news bias?

Core argument: An analysis of over one million TV transcripts finds that coverage of gasoline prices ramps up sharply once the nominal price hits $3.50 per gallon, which the authors characterize as a negative news bias.

We analyzed over one million transcripts and found that TV coverage of gas prices exhibits a negative news bias. TV coverage sharply ramps up when the nominal gas price hits $3.50 per gallon. The ramp-up is steeper for cable than network outlets, and among cable outlets, steepest for Fox News. Because the $3.50 threshold is stable in nominal terms, the real threshold at which the media ramps up their coverage of gas prices has decreased by one-third over time.

Takeaways by Macro Roundup® AI

  1. An analysis of over one million TV transcripts finds that coverage of gasoline prices ramps up sharply once the nominal price hits $3.50 per gallon, which the authors characterize as a negative news bias.
  2. The ramp-up in gas price coverage at the $3.50 nominal threshold is steeper for cable than for network outlets, and steepest among cable outlets for Fox News.
  3. Because the $3.50 threshold is stable in nominal terms, the real gas price threshold at which the media ramps up coverage has decreased by one-third over time.

Related Articles:

  • Americans Are Suddenly a Lot More Upbeat About the Economy — The University of Michigan's consumer sentiment survey recorded a 29% rise over two months, the largest two-month increase since 1991. The improvement was broad-based across consumers of different ages, incomes, education levels, and regions.
  • The Cost of Money is Part of the Cost of Living: New Evidence on the Consumer Sentiment Anomaly — Variation in the University of Michigan consumer sentiment index that inflation and unemployment cannot explain is strongly correlated with growth in consumer borrowing costs, such as mortgage and car loan rates. The authors estimate that including interest payments and homeownership costs in the consumer price index would raise measured inflation from 3% to 9%.

AI Summary. Encounters between migrants and US immigration authorities fell to about 190,000 per month in February and March from a record of more than 300,000 in December, after Mexico restored funding for migration services and increased deportations. Administration officials say encounters over four months were down 40% from the prior four-month period.

Eric Martin Bloomberg
Date Posted:

CBP data show border encounters fell from 300,000 in December to about 190,000 monthly in the first four months of 2024 after Mexico stepped up enforcement measures.

Have regional enforcement efforts reduced migrant encounters at the southern border?

Core argument: Migrant encounters at the US southern border in the first four months of this year fell 40% versus the final four months of last year, according to senior administration officials speaking on condition of anonymity.

The number of encounters between migrants and immigration authorities during the first four months of this year is down 40% compared with the final four months of last year, according to senior administration officials, who spoke on the condition they not be identified. The decline implies that the number of arrivals by undocumented migrants has continued to slide from March into April, the first time during the Biden administration that the figures slowed during that span despite warmer weather, when the figures would typically rise. Customs and Border Protection data show that border encounters fell to about 190,000 per month in February and March, down from a record of more than 300,000 in December, after Mexico restored funding for migration services and ramped up deportations.

Takeaways by Macro Roundup® AI

  1. Migrant encounters at the US southern border in the first four months of this year fell 40% versus the final four months of last year, according to senior administration officials speaking on condition of anonymity.
  2. Customs and Border Protection data show border encounters dropped to about 190,000 per month in February and March from a record of more than 300,000 in December, after Mexico restored funding for migration services and increased deportations.
  3. The decline implies undocumented arrivals kept falling from March into April, the first slowdown over that span during the Biden administration despite warmer weather, when the figures would typically rise.

Related Articles:

  • Illegal US Border Crossings Aren’t Really Breaking Records — Official US border encounter statistics exclude people who evade Border Patrol agents, who in past decades made up the majority of illegal crossers. The Department of Homeland Security estimated apprehension rates of 33%–43% in the 2000s versus 81% more recently, a shift attributed mainly to most crossers intending to turn themselves in and apply for asylum.
  • Immigration and the U.S. Economy Since the Pandemic: An Accounting Exercise — 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.
  • Immigration’s Effect on US Wages and Employment Redux — The authors estimate that immigration to the US over 2000-2019 raised wages of natives with a high school degree or less by 1.7% to 2.6%, a period when that group's real wages fell by 5% to 6%. Effects on more-educated natives ranged from -0.5% to +0.7%, with an overall average effect of +0.5% to +0.8% that was not statistically significant.

AI Summary. In a randomized trial of 15,965 hospitalized patients, an artificial-intelligence electrocardiogram system that flagged patients predicted to be at high risk of death and sent warning messages to physicians was associated with lower 90-day all-cause mortality, 3.6% in the intervention group versus 4.3% in the control group.

Chin-Sheng Lin, Wei-Ting Liu, Dung-Jang Tsa, et al. Nature Medicine
Date Posted:

A study of 15,965 cardiac patients in Taiwan found that AI-enabled electrocardiogram alerts to physicians reduced the 90-day mortality of high-risk patients from 4.3% in a control group to 3.6% in the AI-ECG group.

Do AI risk alerts sent to physicians reduce patient mortality?

Core argument: In a multisite randomized controlled trial, an AI-enabled ECG alert was associated with 90-day all-cause mortality of 3.6% in the intervention group versus 4.3% in the control group, meeting the trial's primary outcome.

This study evaluated the ability of an artificial intelligence (AI)-enabled electrocardiogram (ECG) to identify hospitalized patients with a high risk of mortality in a multisite randomized controlled trial involving 39 physicians and 15,965 patients. The AI-ECG alert intervention included an AI report and warning messages delivered to the physicians, flagging patients predicted to be at high risk of mortality. The trial met its primary outcome, finding that implementation of the AI-ECG alert was associated with a significant reduction in all-cause mortality within 90 days: 3.6% of patients in the intervention group died within 90 days, compared to 4.3% in the control group (4.3%).

Takeaways by Macro Roundup® AI

  1. In a multisite randomized controlled trial, an AI-enabled ECG alert was associated with 90-day all-cause mortality of 3.6% in the intervention group versus 4.3% in the control group, meeting the trial's primary outcome.
  2. The trial involved 15,965 hospitalized patients and 39 physicians, with the intervention delivering an AI report and warning messages flagging patients predicted to be at high risk of mortality.

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.

Friday, May 10, 2024

AI Summary. A scenario analysis projects that the U.S. public corporate sector's overall debt-servicing capacity is robust to sustained elevated interest rates, aided by low borrowing costs on existing debt and moderate near-term refinancing needs. Firms with weaker balance sheets, including some large investment-grade borrowers, are projected to see meaningful credit-quality deterioration, with the share of debt at risk rising.

Dalida Kadyrzhanova, Ander Perez-Orive, and Eliezer Singer Federal Reserve Board
Date Posted:

A @FederalReserve analysis of four different scenarios finds that absent a significant recession, the debt servicing capacity of US non-financial corporations is strong despite higher rates and an 8pp rise in debt as a % of GDP.

Can corporate borrowers withstand sustained high interest rates?

Core argument: The scenario analysis finds the U.S. public corporate sector's debt-servicing capacity robust to sustained elevated interest rates in both a soft-landing baseline and a stagflation scenario with a moderate downturn, but not in a severely adverse recession scenario.

The debt-servicing capacity of the U.S. public corporate sector as a whole is robust to sustained elevated interest rates, both in the soft landing (baseline) scenario as well as in a stagflation scenario with a moderate economic downturn [though not the “severely adverse” recession scenario]. This resiliency is due, in large part, to the strong balance sheets of public firms—namely, low borrowing costs on existing debt and moderate refinancing needs in the short run. In the distribution of firms, however, persistently higher interest rates are projected to result in a meaningful deterioration in the credit quality of firms with weaker balance sheets, including some large investment-grade firms that have so far been relatively insulated from rising rates by their high share of fixed-rate debt. Accordingly, the share of debt at risk is likely to see sustained increase over the next two to three years even if earnings remain resilient.

Takeaways by Macro Roundup® AI

  1. The scenario analysis finds the U.S. public corporate sector's debt-servicing capacity robust to sustained elevated interest rates in both a soft-landing baseline and a stagflation scenario with a moderate downturn, but not in a severely adverse recession scenario.
  2. The analysis attributes corporate resilience largely to strong balance sheets at public firms, specifically low borrowing costs on existing debt and moderate refinancing needs in the short run.
  3. Persistently higher interest rates are projected to meaningfully deteriorate credit quality among firms with weaker balance sheets, including some large investment-grade firms so far insulated by high fixed-rate debt shares, with the share of debt at risk rising over two to three years even if earnings remain resilient.

Related Articles:

  • The Corporate Debt Maturity Wall: Implications for Capex and Employment — The authors estimate the average interest rate on outstanding corporate debt will rise from 4.20% to 4.50%, lifting private sector interest expense as a share of gross output by 0.25 percentage points. They project this would reduce capital spending growth by up to 0.25 percentage points and labor cost growth by up to 0.15.
  • Corporate Debt, Boom-Bust Cycles, and Financial Crises — Using a dataset on sectoral credit exposure for 115 economies, the authors find corporate debt accounts for two-thirds of credit expansion before financial crises and three-quarters of nonperforming loans during busts, and that corporate debt expansions predict crises similarly to household debt. Recoveries are slower after corporate debt booms.
  • Rates Are Up. We’re Just Starting to Feel the Heat — U.S. companies have $600bn in corporate debt maturing this year, a total Goldman Sachs data show rising to more than $1tn a year from 2025 through 2028. Those debt loads, combined with higher financing costs, may cut into corporate profits, investor returns, innovation spending, and hiring, with some analysts expecting credit-rating downgrades.

AI Summary. Bank of America internal payments data show average household property insurance payments up 23% relative to 2019, and average per-customer payments up 25%, with a three-month average year-over-year increase of 11%. The data cannot separate auto from home insurance payments.

David Michael Tinsley, Joe Wadford, et al. Bank of America
Date Posted:
Is Database:
Database

An analysis of @BankofAmerica internal data indicates that the mean monthly customer payment for home and auto insurance is up 25% since 2019, and represented around 4.5% of household income in the 12 months to March 2024.

How much have household property insurance payments risen?

Core argument: Bank of America internal payments data show the average property insurance payment is up 23% relative to 2019, with the average customer payment up 25%, a gap reflecting customers making more than one insurance payment in a month.

We use Bank of America internal data across payment channels to examine how households’ overall property insurance payments are changing. While we cannot distinguish in the data between auto or home insurance payments, we can use our data to gauge the impact of rising aggregate property insurance costs across different groups of consumers. The average insurance payment is up 23% relative to 2019 and the average customer payment is up 25%. The difference between these two numbers reflects the fact that some customers will make more than one insurance payment in any month. Compared with 2023, our data shows a rapid increase in these costs for customers over the past year – with the three-month average April 2024 YoY rise at 11%.

Takeaways by Macro Roundup® AI

  1. Bank of America internal payments data show the average property insurance payment is up 23% relative to 2019, with the average customer payment up 25%, a gap reflecting customers making more than one insurance payment in a month.
  2. Bank of America data show property insurance costs rising rapidly for customers over the past year, with the three-month average April 2024 year-over-year increase at 11%.
  3. Bank of America's payments data cannot distinguish auto from home insurance payments, but can gauge the impact of rising aggregate property insurance costs across different groups of consumers.

Related Articles:

  • As Texas Burns, Property Insurance Costs Soar — Rising property insurance premiums are attributed to climate-related damage, higher rebuilding costs, and costlier reinsurance protection for insurers, which is passed on to homeowners. The insurance industry says rebuilding and replacement costs rose 55% over a three-year span, driven by raw materials, wages, supply-chain bottlenecks, and a rising share of larger homes.
  • In Southwest Florida, High Home Insurance Rates Are Driving Away Would-Be Buyers — Florida homeowners paid about $6,000 a year for insurance on average, roughly triple what they paid in 2019 and compared with about $1,700 for the average US homeowner. Insurers and state regulators attribute the increases to widespread lawsuits and fraud, and a Category 5 hurricane led some insurers to exit the state or limit new policies.
  • A Hidden Crisis In US Housing — State-run "last resort" home insurance programs, which cover properties the private market won't, have more than doubled their market share since 2018, with liabilities passing $1tn; Florida's plan faces potential losses of $525bn and California's at least $290bn, six times its 2018 level. Of 36 such plans covering natural catastrophes, 21 do not explicitly detail how they would pay deficits.

AI Summary. California has installed more grid-scale batteries than any place outside China, storing excess daytime solar power for use after dark and supplying over one-fifth of the state's electricity during some evening hours. National battery storage capacity has grown tenfold over three years to 16,000 megawatts and is expected to nearly double again, led by Texas, California, and Arizona.

Brad Plumer and Nadja Popovich New York Times
Date Posted:
Is Database:
Database

Grid-scale batteries can supply up to 20% of California’s electrical demand, equivalent to the output of 7 large nuclear reactors. US battery storage has grown 10X in the past three years to 16,000 megawatts and is expected to double again this year.

Are giant batteries changing how the US stores and uses electricity?

Core argument: U.S. grid battery storage capacity has grown tenfold over the past three years to 16,000 megawatts and is expected to nearly double again this year, with the biggest growth in Texas, California, and Arizona.

Since 2020, California has installed more giant batteries than anywhere in the world apart from China. They can soak up excess solar power during the day and store it for use when it gets dark. Those batteries play a pivotal role in California’s electric grid, partially replacing fossil fuels in the evening. Between 7 p.m. and 10 p.m. on April 30, for example, batteries supplied more than one-fifth of California’s electricity and, for a few minutes, pumped out 7,046 megawatts of electricity, akin to the output from seven large nuclear reactors. Across the country, power companies are increasingly using giant batteries the size of shipping containers to address renewable energy’s biggest weakness: the fact that the wind and sun aren’t always available. Over the past three years, battery storage capacity on the nation’s grids has grown tenfold, to 16,000 megawatts. This year, it is expected to nearly double again, with the biggest growth in Texas, California, and Arizona.

Takeaways by Macro Roundup® AI

  1. U.S. grid battery storage capacity has grown tenfold over the past three years to 16,000 megawatts and is expected to nearly double again this year, with the biggest growth in Texas, California, and Arizona.
  2. Between 7 p.m. and 10 p.m. on April 30, batteries supplied more than one-fifth of California's electricity, peaking for a few minutes at 7,046 megawatts, akin to the output of seven large nuclear reactors.
  3. California has installed more giant batteries since 2020 than anywhere in the world apart from China, soaking up excess daytime solar power and partially replacing fossil fuels in the evening.

Related Articles:

  • U.S. Battery Storage Capacity Expected To Nearly Double In 2024 — U.S. battery storage capacity is projected to rise 89%, to more than 30 gigawatts, if developers bring all planned energy storage systems online as scheduled — a total that would exceed capacity from petroleum liquids, geothermal, wood waste, or landfill gas. California leads all states with 7.3 gigawatts installed, followed by Texas with 3.2 gigawatts.
  • 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.
  • Electricity Grids and Secure Energy Transitions — Under a scenario based on announced government pledges, global electricity demand is projected to grow 2.7% per year, more than doubling from just under 25,000 terawatt-hours to nearly 54,000 terawatt-hours by 2050. Hydrogen production via electrolysis and electric vehicles account for more than half of that projected growth.

AI Summary. AlphaFold 3, a single deep learning model with a diffusion-based architecture, jointly predicts the structures of complexes containing proteins, nucleic acids, small molecules, ions, and modified residues. The authors report accuracy exceeding specialised tools for protein-ligand docking, protein-nucleic acid interactions, and antibody-antigen prediction.

Josh Abramson, Jonas Adler, Jack Dunger, et al. Nature
Date Posted:

.@GoogleDeepMind and @IsomorphicLabs announced AlphaFold 3, claiming a 50% accuracy improvement in the interaction between proteins, DNA, RNA, and other molecules. “We hope it will transform our understanding of the biological world and drug discovery.”

Can a single deep learning model predict diverse biomolecular structures accurately?

Core argument: AlphaFold 3 uses a substantially updated diffusion-based architecture to jointly predict structures of complexes spanning proteins, nucleic acids, small molecules, ions, and modified residues.

The introduction of AlphaFold 2 [July 2021] has spurred a revolution in modelling the structure of proteins and their interactions, enabling a huge range of applications in protein modelling and design. In this paper, we describe our AlphaFold 3 model with a substantially updated diffusion-based architecture, which is capable of joint structure prediction of complexes including proteins, nucleic acids, small molecules, ions, and modified residues. The new AlphaFold model demonstrates significantly improved accuracy over many previous specialised tools: far greater accuracy on protein-ligand interactions than state of the art docking tools, much higher accuracy on protein-nucleic acid interactions than nucleic-acid-specific predictors, and significantly higher antibody-antigen prediction accuracy. Together these results show that high accuracy modelling across biomolecular space is possible within a single unified deep learning framework.

Takeaways by Macro Roundup® AI

  1. AlphaFold 3 uses a substantially updated diffusion-based architecture to jointly predict structures of complexes spanning proteins, nucleic acids, small molecules, ions, and modified residues.
  2. The authors report that AlphaFold 3 achieves far greater accuracy on protein-ligand interactions than state-of-the-art docking tools, much higher accuracy on protein-nucleic acid interactions than nucleic-acid-specific predictors, and significantly higher antibody-antigen prediction accuracy.
  3. The authors conclude that high-accuracy modelling across biomolecular space is possible within a single unified deep learning framework.

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.