How to measure the returns on R&D spending

Recent economic research finds public and basic R&D delivers unusually high long-term returns, but proposed federal cuts threaten future productivity and areas such as Artificial Intelligence and quantum information science.

Debate over how much the government should invest in research and development has intensified as the administration proposed steep reductions to major science agencies. The article frames the question as one of return on investment. While many celebrated technologies trace back to public funding, counting successes alone gives a misleading picture. Returns from research are hard to measure because payoffs often take years and travel indirect paths, and many projects fail to produce usable outcomes.

Several recent papers provide fresh, systematic evidence that R&D is among the better long-term public investments. Benjamin Jones and Lawrence Summers calculate large social returns to national R&D spending, though the article notes key numeric assumptions are not stated. Andrew Fieldhouse and Karel Mertens trace changes in public R&D at five major agencies and estimate that nondefense public R&D has driven roughly 20 to 25 percent of private-sector productivity growth in the United States since World War II. The economists find benefits typically emerge after about five to ten years and persist for decades. The piece also records that federal R&D peaked at 1.86 percent of GDP in 1964 and has fallen to about 0.6 percent of GDP, while business R&D has risen.

Other work highlights the special role of public funding for fundamental science. Arnaud Dyèvre finds public R&D spillovers produce about three times the productivity impact of private R&D, suggesting the country may be underinvesting in basic research. The article flags specific policy risks: the administration proposed cutting the National Institutes of Health budget by 40 percent and the National Science Foundation by 57 percent, and some agency directorates that support fields critical to advancing Artificial Intelligence and quantum information science face even larger proportional cuts. Economists quoted warn such reductions would measurably slow productivity growth over the next seven to ten years. A new five-year research effort funded by foundations will try to refine estimates of optimal R&D spending. The bottom line in the article is clear: recent empirical work strengthens the case that more public funding for basic science is a sound investment in long-term prosperity.

70

Impact Score

Saudi Artificial Intelligence startup launches Arabic LLM

Misraj Artificial Intelligence unveiled Kawn, an Arabic large language model, at AWS re:Invent and launched Workforces, a platform for creating and managing Artificial Intelligence agents for enterprises and public institutions.

Introducing Mistral 3: open artificial intelligence models

Mistral 3 is a family of open, multimodal and multilingual Artificial Intelligence models that includes three Ministral edge models and a sparse Mistral Large 3 trained with 41B active and 675B total parameters, released under the Apache 2.0 license.

NVIDIA and Mistral Artificial Intelligence partner to accelerate new family of open models

NVIDIA and Mistral Artificial Intelligence announced a partnership to optimize the Mistral 3 family of open-source multilingual, multimodal models across NVIDIA supercomputing and edge platforms. The collaboration highlights Mistral Large 3, a mixture-of-experts model designed to improve efficiency and accuracy for enterprise artificial intelligence deployments starting Tuesday, Dec. 2.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.