🧪⚛🧬 Broken promises: CHIPS Act funding for science research falls short
More R&D spending by Washington should be a no-brainer. But it's not.
Quote of the Issue
“The emergence in the early twenty-first century of a new form of intelligence on Earth that can compete with, and ultimately significantly exceed, human intelligence will be a development of greater import than any of the events that have shaped human history.” - Ray Kurzweil, The Age of Spiritual Machines: When Computers Exceed Human Intelligence
The Conservative Futurist: How To Create the Sci-Fi World We Were Promised
“With groundbreaking ideas and sharp analysis, James Pethokoukis provides a detailed roadmap to a fantastic future filled with incredible progress and prosperity that is both optimistic and realistic.”
The Essay
🧪⚛🧬 Broken promises: CHIPS Act funding for science research falls short
Anyone who has a future-optimist, techno-solutionist, Up Wing outlook will almost surely find the following two charts incredibly aggravating — if not outright demoralizing:
The clear message of those two charts: Despite the promises made in the CHIPS and Science Act to boost funding for scientific research and development in the US, Congress has significantly underfunded key agencies and programs as intended by the 2022 law. As Politico reports:
While nearly $53 billion is going into reviving a homegrown semiconductor industry, Congress has gnawed away at the law’s ambitions on fundamental research and development aimed at staying ahead of China and other rivals in competitive fields like artificial intelligence. The latest example is the spending package lawmakers advanced over the past week: Biden’s signature enacts deep cuts to the National Science Foundation and stalls key offices in the Commerce and Energy departments that are supposed to deploy CHIPS money, turning a promised cash infusion of $200 billion over a decade into a humiliating haircut. And while it’s hardly the first time Congress has reneged on promised funds, it threatens an important pillar of Biden’s industrial policy. … The National Science Foundation offers a window onto the broader challenge, as the agency that expected to gain the most from CHIPS. The law authorized NSF to receive $81 billion over a five-year period that would ultimately double the agency’s budget by 2027. This year, the NSF should have gotten $15.64 billion, according to the CHIPS Act. Under the latest spending deal, lawmakers provided NSF with $9.06 billion — 42 percent short of the CHIPS target and an 8 percent cut to its current budget.
While the above piece is correct that increased science and technology research is a part of “Biden’s industrial policy” in a broad, non-technical sense, economists would not typically classify such investment as industrial policy, per se. Basic research to expand fundamental knowledge, and even perhaps applied research to solve practical problems or develop new technologies, isn’t meant to develop specific products or technologies for commercial use. The knowledge created by basic research, in particular, is often considered a public good, meaning that it is non-excludable and non-rivalrous. This makes it difficult for private companies to fully capture the benefits of basic research, leading to underinvestment and the need for government support.
Industrial policy, on the other hand, involves subsidies, tax incentives, trade barriers, and other targeted interventions designed to promote the competitiveness of specific industries. The CHIPS part of the CHIPS and Science Act, meant to boost domestic semiconductor production, is a more classically industrial policy.
Here’s another irritating chart, showing federally funded R&D support as a share of GDP at approximately half its level from the 1980s and more generally is at its lowest level since the Space Age and Project Apollo (via the AAAS):
Consider: The next time you take an Uber or Lyft, take a moment to think about Albert Einstein. The physicist’s famous theories of relativity, which explain the effects of velocity and gravity on time, were essential for the development of GPS technology. GPS, in turn, enabled the creation of smartphones and novel business models like those of ride-sharing companies.
How did we get from Einstein’s theories in 1905 to Uber’s launch in 2009? Einstein's theories are essential for making the necessary adjustments to the atomic clocks on GPS satellites, ensuring their accuracy. (These satellites orbit the Earth at high velocities and experience less gravity compared to users on the ground, meaning time ticks at a different rate on the satellites, as predicted by Einstein.) Uber, in turn, depends on this GPS technology to find you and get you efficiently where you want to go.
This series of what economists term “ intertemporal spillovers” — where one scientific advance unlocks doorways to further innovation — is a key source of “social value,” or the broader and cascading benefits that scientific and technological advances bring to society as a whole, beyond the immediate financial gains for the innovative scientists and technologists.
Another example: In 1969, University of Indiana biologists Thomas Brock and Hudson Freeze discovered Thermus aquaticus, a bacterium that thrives in Yellowstone National Park's hot springs. This discovery had no immediate practical application. But in the 1980s, biochemist Kary Mullis was searching for an enzyme to efficiently replicate human DNA at high heat, where DNA unwinds and can be copied. Mullis recalled Thermus aquaticus, realizing it must replicate its DNA at high heat. The bacterium's replication enzyme revolutionized the biotechnology industry by making possible polymerase chain reaction, or PCR, the advance that enables genetic testing, cancer diagnosis, gene-based drug development, and more. Mullis received a Nobel Prize for this work, while we get — among those other things — daytime reality television, with its many dramatic paternity reveals.
The theories of relativity and the discovery of the thermophile bacterium are both examples of the unpredictable, long-run benefits of basic science, ones highlighted by economist Benjamin Jones in his excellent 2021 analysis “Science and Innovation: The Under-Fueled Engine of Prosperity.” What magnitude of benefits are we talking about? From that paper,
Looking purely at the social returns, the standard findings suggest that doubling the total investment in R&D would easily pay for itself. That is, the additional expansion in standards of living in terms of GDP per person would be much larger in present value than the additional investment cost. How much potential is the United States leaving on the table? Using the general approach in Jones and Summers, a sustained doubling of all forms of R&D expenditure in the U.S. economy could raise U.S. productivity and real per-capita income growth rates by an additional 0.5 percentage points per year over a long time horizon. This would lead to enormous increases in standards of living over time. It would greatly advance the competitiveness of U.S. businesses and workers and the overall position of the U.S. economy in the world. And this economic orientation leaves out the health gains of longer and healthier lives, which are among the most valuable deliverables from the science and innovation system.
Since the late 1960s, the United States has experienced a significant slowdown in productivity growth, with the exception of a short-lived uptick in the late 1990s and early 2000s. Interestingly, this downturn coincides with a reduction in public investments for R&D. But is this merely a correlation, or is there a causal relationship between the two?
In a December paper titled "The Returns to Government R&D: Evidence from U.S. Appropriations Shocks," economists Andrew J. Fieldhouse and Karel Merten from the Dallas Fed sought to answer this question. They examined historical federal budgets to identify significant changes in government research funding approvals, categorizing them as either defense or nondefense. The researchers then analyzed how these funding shocks affected the business sector's Total Factor Productivity over extended periods, up to 15 years after the initial funding changes. (TFP is a measure of economic productivity that accounts for the efficiency and effectiveness of all inputs used in production, including labor, capital, and technology, often used as shorthand for innovativeness.) As they conclude:
Our estimates indicate that government-funded R&D accounts for roughly one quarter of all business sector productivity growth since World War II, including one quarter of the deceleration in productivity growth since the late 1960s. Correlation does not imply causation in general, but our new causal evidence lends support to the thesis of Gruber and Johnson about the important relationship between government-funded R&D and U.S. productivity growth.
In The Conservative Futurist, I advocate returning federal spending to Apollo levels — preferably contingent on Washington funding further research into boosting science productivity, as well as implementation of “metascience” funding reforms — but also by making the corporate tax code more research-friendly.
Nor should we forget that human capital policies are key drivers of innovation. Children exposed to innovators and entrepreneurs, especially girls in regions with more female inventors, are more likely to become inventors themselves. Immigrants also play a crucial role, patenting more often than native-born Americans, making up a significant portion of the science and engineering workforce, and being more likely to start companies of all sizes, including high-growth start-ups.
Let’s make sure the next big sci-tech breakthrough — as with generative AI, reusable rockets, and nuclear fusion — also happens in America, and not China.
Micro Reads
Business and Economics
▶ Dimon Hails ‘Unbelievable’ AI, Says Central to Conversations - Bberg
▶ Populism and the skill content of globalisation - VoxEU
▶ Toward an Understanding of the Returns to Cognitive Skills Across Cohorts - NBER
▶ The future of ‘communist capitalism’ in China - FT Opinion
▶ The Wealthy Are Starting to Have More Babies Than the Poor Again - Bberg Opinion
▶ Edtech Unicorns Are Evolving Rather Than Disrupting - Bberg
Policy
▶ European Lawmakers Pass AI Act, World’s First Comprehensive AI Law - WSJ
▶ Permitting the Energy Transition - Case Western Reserve Law Review
▶ Biden seeks to boost science funding — but his budget faces an ominous future - Nature
▶ When It Comes to Big Tech, Regulatory Ambition Ignores Consumers’ Choices - AEIdeas
AI
▶ Algorithmic progress in language models - Arxiv
▶ Large Language Models Can Enhance Persuasion Through Linguistic Feature Alignment - SSRN
▶ An AI that can play Goat Simulator is a step toward more useful machines - MIT Tech Review
▶ Arm unveils first chip design to power self-driving cars - FT
▶ Don’t Fear AI in War, Fear Autonomous Weapons - Bberg
▶ Could an AI replace all music ever recorded with Taylor Swift covers? - NS
▶ Your Kid May Already Be Watching AI-Generated Videos on YouTube - Wired
▶ Google’s Bad Gemini Rollout Did the World a Favor - Bberg Opinion
▶ Automakers Are Sharing Consumers’ Driving Behavior With Insurance Companies - NYT
▶ China’s Best Self-Driving Car Platforms, Tested and Compared - Wired
Health
▶ Why are so many young people getting cancer? What the data say - Nature
▶ Global Ozempic shortage fix: Cheaper, faster method produces 10x more - New Atlas
▶ Some states are now trying to ban lab-grown meat - Ars
▶ The Gift of a Lifetime: The Hospital, Modern Medicine, and Mortality - AEA
Cities
▶ Vancouver’s new mega-development is big, ambitious and undeniably Indigenous - Macleans
Clean Energy
▶ How to Feed the AIs - Casey Handmer
▶ Solar Success Is a Curse for China’s Manufacturers - Bberg Opinion
Robotics
▶ Dyson’s 360 Vis Nav is a powerful robovac, but it’s the brush that tickles me the most - Verge
▶ Covariant Announces a Universal AI Platform for Robots - IEEE
Space/Transportation
▶ Starship launch 3: What time is the SpaceX flight and what to expect - NS
▶ Mining helium-3 on the Moon has been talked about forever—now a company will try - Ars
▶ Varda Hopes New Research Draws More Drugmakers to Space Factories - WSJ
▶ Rocket Explodes Moments After Japan Space Startup’s Launch - Bberg
▶ How quickly is the Universe disappearing from our reach? - Big Think
Up Wing/Down Wing
▶ Ultimate Guide to “AI Existential Risk” Ecosystem - AI Panic
Im confused by the R&D chart as a share of GDP. There are three lines, none labeled. You say this funding is 1/2 of the 1980s, but not one of the lines falls by 50%.