The public funding of research and development

This is one of the best of all government programs (or it can be viewed as a collection of programs).  Here is a good survey of the issue, by Brad Plumer, excerpt:

There’s a long, long list of world-changing innovations that can be traced back to federally funded R&D over the years. The Department of Energy’s research labs spawned digital recording technology, communications satellites, and water-purification techniques. Pentagon research laid the groundwork for the Internet and GPS. The current shale-gas fracking boom couldn’t have happened without microseismic imaging techniques that were developed at Sandia National Laboratories.

It also can be said that this is probably the worst side of the sequester.


What's that they say about never biting the hand that feeds....... :)

I am surprised that you found this persuasive. How much of this funding was wasted? Trying to figure out the overall ROI for government R&D would not be easy and would have to factor in the losers and the dollar value of private initiative in converting an idea into a profitable business. This report seems like those commercials for the NCAA that are shown during football or basketball games.

What is wasted funding? In fundamental research you are always in an uncharted territory, you have to explore many ways most of which won't lead to anything but you have to do that anyway. It's easy to be a general after the battle. You can't solve a math/physics problem by ramping up marketing instead.

The Brad Plumer article linked in the post contains some information and links to studies purporting to show examples of wasted research money, or that research money spent by government is wasted in general. Liberals might argue that defense research on things like missile defense is wasted. Conservatives might argue that research into liberal cuases like green energy or climate change are wasted. It seems unlikely that research spending is unique among all government programs in that it is immune from the waste associated with capture by special interests and rent seekers.

And if DOE hadn't soaked up all those engineers and scientists with unlimited budgets, easy hours, and soft accountability, what would private actors have accomplished?

And if the Feds hadn't appointed aged establishment types to dole out vast sums to, usually, their orthodox acolytes, what would the heterodox, competitive private actors have accomplished?

Maybe nothing, sure, but let's not get all moist over this program without thinking about that.

They probably would have gone into finance to make fancy derivatives and other financial instruments that screw over the global economy a good point. It even holds up if you rephrase it as, "They probably would have gone into [polisci, law, politics] to make fancy [regulations] and other financial [incentives] that screw over the global economy."

Then you're not talking about "engineers and scientists", because they wouldn't have gone into law or politics. But they would maybe have gone into finance.

There does seem to be a lack of engineers in politics, but certainly not law.

Many engineers enter the field of law due to a complicated patent system, or liabilities associated with a product. Like a defective vehicle, or a levy that needs maintenance. Do not write them out of that field.

None of the (NIH or NSF funded) scientists that I know have anything close to even a single one of "unlimited budgets, easy hours, and soft accountability".

"easy hours, and soft accountability”

Lots of NIH or NSF funded scientist have easy hours (not all or even majority, of course) and practically everyone has soft accountability.

And if DOE hadn’t soaked up all those engineers and scientists with unlimited budgets

You're very funny.

This is the old Washington monument scenario. Minor cuts whose implementation is designed to effect only the expenditures most useful or noticed.

I'd suggest boiling up a pot of tar and collecting a bunch of chickens. Or putting to rest the idea that the democrats are fiscally responsible.

This is one of the best of all government programs

That's what every agency and organization impacted by any budgets cuts says.

Defending our country / defense programs: This is one of the best of all government programs

Educating our children: This is one of the best of all government programs

Food stamps / feeding the hungry and needy among us: This is one of the best of all government programs

Building and maintaining highways: This is one of the best of all government programs

The DEA and Drug War: This is one of the best of all government programs

Supporting family farms / ag subsidies: This is one of the best of all government programs

No one likes it when their favorite project or program that spends other people's money gets cut back or axed.

Though I hate technological just-so stories most of all, I'll have to take the economist's word for it. There is something to be said for the government seeing fit to provide one of its legitimate roles, even though it has screwed itself out of that ability.

There's no question that public R&D had beneficial outcomes, produced public goods, and had positive externalities. But as Rich says we haven't observed the costs and ROI. We aren't considering the failed projects or negative NPV projects and the government's track record in decision making.

The number of admissible public projects is enormous. A private firm would have to maximize NPV subject to a budget constraint. A government with the capacity to undertake projects without a budget constraint is the road to fiscal ruin.

One need only look at World War I, World War II and the Cold War for examples of public goods and R&D that busted national budgets.

Givco also raises the excellent point of crowding out. The US government didn't create the automotive airbag, it just mandated it. Same goes for employment based health insurance.

Think of all the weapon R&D that took place as a result of the high capacity magazine ban. It unexpectedly ushered in entire classes of concealable ten round large caliber pistols.

Why is it an indictment of government R&D spending when a government-funded project fails but healthy creative destruction when it's a privately funded project? Both are wealth expenditures that netted society nothing. I get that with private R&D it's not your money being lost, but with private R&D, the benefits are also privatized to a large degree. If you think the government's current levels of R&D investment is crowding out the private sector, you must think there's not a lot of things out there worth investing research time/funds in.

Market discipline is.

A 4% reduction in funding is going to have an impact? You could probably cover that just by firing unnecessary staff. Also, why does R&D have to continuously increase as a percentage of GDP? It has to taper off at some point.

Yes, that is the most moronic defend funding for science I've ever seen.

Instead of acknowledging no system, even science funding system, is perfect. And recognizing that some improvements, like funding cuts, can be done without affecting final results; the debate degenerated into who's better for doing "science".

The debate is unnecessary, government funds basic research that (just a guess) 5% success ratio, while private companies fund commercial research with a (just a guess) 50% success ratio. And that is fine, basic research done 100 years ago is used to do commercial research right now.

Using computers as example. Textile makers had punched cards, so? They couldn't accomplish a simple 1+1 operation. Some guy at Columbia U applied the punched card to do some calculations, later worked for US Census and funded a company that later mutated into IBM. Some other people constructed the first electronic computer to calculate artillery trajectories during WWI. That job needed SOFTWARE to work. Government paid that computer and the software. Later private companies started to do electronic hardware and standard software like Fortran and C++. Right now software is written using the math that people with long and complicate last names produced a century or more ago. And whatever people say, private companies do not do math research So, goverment (universities) and private companies are great partners, not antagonists.

Perhaps the 4% goal is achievable by firing some people from NSF and decomissioning old research facilities.

Information is obviously a public good. In general private actors will not have sufficient incentive to produce information since they do not internalize all of the benefits of the information once that information is available. Of course, to the extent possible they would just not allow others access to that information. But that seems like a bad outcome to me. For instance, if a private company discovers a more efficient way to manufacture steel, do we really want that information to *not* be publicly available?

Companies don't internalize all of the benefits of *anything* they do. And as for your last sentence, there are ways to make Hank Rearden talk.

I notice Tyler didn't include an example from social science. Has Tyler written about Cantor's efforts to get the NSF to stop funding political science? If so I seem to have missed it.

It has been brought up. I was about to say above that perhaps some portion of public research (and we already have this to some extent) should be sponsored by private interests. However that introduces problems. The flip side of the coin is government using private money (taxes and bonds) to fund political science.

"It also can be said that this is probably the worst side of the sequester. "

The sequester just isn't that big. If these cuts on the order of 4% real cuts are a huge deal, what are to make of the scheduled 20+% Medicaid reimbursement cuts?

Government funded research invented the jet engine, the computer, the very large scale integrated circuit and the Internet. Those projects more than paid for every research failure that the government might ever have suffered.

That's statement is so wrong, I have to ask is it parody?

" is it parody"

No. Government clearly paid for the development of jet engines (through the German, English and American military), government paid for Eniac, DARPA paid for the original development of the Internet (CERN, which is also government funded, paid for the invention of the Web) and DARPA paid for the VLSI project which made modern (post 1980) microprocessors possible.

I think he was referring to your second sentence; Given the nature of much government research, neither of you could possibly know whether that statement is right or wrong. FWIW, government funding did not pay for the invention or development of the jet engine.

Yes, that was my primary point. It's a completely unprovable assertion.

But his statement was worse than just that.

1) As you noted government funding didn't pay for the invention of the jet engine:

2) Government funding didn't pay for the invention of the computer. Indeed, the first computers were created in the early 19th century for the textile industry.

3) Jack Kilby at Texas Instruments created the first integrated circuits in 1958. A whole line of successfully more advanced processors were developed in the 1960's and 1970's. And while Darpa did fund the VLSI project, that was merely a refined version of a product that had already been developed by private industry. The first VLSI chips predate Darpa funding of the VLSI project.

To quote: "DARPA's VLSI Project provided research funding to a wide variety of university-based teams in an effort to improve the state of the art in microprocessor design, then known as VLSI.

To be fair, I would agree that the development of the Internet was funded by the US military.

I said the development not invention of the jet engine was paid for by government. Taking the jet engine from concept to practical product between the late thirties and early 50s was a government project.

The von Neuman machine digital computer came out of the government founded Eniac project.

Integrated circuits were invented around 1960 and the microprocessor was developed around 1970. The Darpa project made possible a higer level of complexity that allowed the microprocessors we have now.

The economic value of jet engines, the internet and modern microprocessors would be measured in the multiple trillions of dollars. That should be more than enough to pay for all unsuccessful government funded research.

"Government funded research invented the jet engine"

-Joe Smith, February 26, 2013 at 4:18 pm

Development of the jet engine was not a government funded project. It was funded by industry that hoped to, among other things, sell to the government, and was quite successful doing so. The government likes jet engines, and has bought a lot of them, but saying they were developed as a government project is like saying the Crown Vic was developed as a government project.

No, but yours is.

These were byproducts of war. Better to have the byproducts or not to have the wars?

Yeah, thank God for the V-2 rocket. Funny stuff. But seriously, first he means "government directed" research. The government funds almost nothing.

Btw, economists, isn't the distinction between public and private provision of R&D NOT about the successes? It's not even about the success rates, right? It's not even about the areas of research, right? It's about what would be underpursued by private interests, right?

The US DOE (and affiliated agencies) also gave billions spent on breeder reactors, synfuels, and an untold number of years of investment into "clean coal technology." To evaluate the success rate of speculative research projects, we should consider the portfolio rather than a handful of big winners.

To further evaluate claims about the success of government research, we should (as the attached article mentions, but not in the clip Tyler has included) consider the appropriate counterfactual: what research and other benefits would have been achieved in the absence of government spending on these research projects.

Proponents of government-research (i.e. taxpayer-funded, politically-attuned, bureaucratically-directed research) should at least show that average research value is at least as great as produced by the likely counterfactual situation.

A very reasonable follow-up discussion concerns whether or not government research policy is focused on maximizing the public good quality of its projects. When government finances a company's investment in developing its own projects (i.e. Solyndra, to use the obvious case), the technological development belongs to the company in ways that seem to reduce the positive externality that research proponents speak of.

While there may be many cases for which good reasons dictate keeping research results secret, the default case should be that the product of government-sponsored research be freely and publicly available.

You shouldn't forget that the government (or other governments) also funded the successful PWR / BWR reactors and probably most of the physics, chemistry and engineering that had to be done decades or centuries before to make them happen.

Same with modern electronics. It's all based on publicly funded (mostly) obscure research that was probably at the time thought by some to be wasteful because it didn't lead to massive profits that same fiscal year.

"Synfuels" is a pretty broad topic. The U.S. military, for instance, is extremely interested -- for reasons that ought to be obvious -- in having the capability of quickly switching to alternative fuels to operate tanks and bombers if it needs to. America's close allies in Central and Eastern Europe are also heavily dependent at the moment on oil and natural gas imports from Russia and it's safe to say that Russia is quite aware of this fact. Research on synthetic fuels appears to be very much a public good given its national security and global security implications. Some of this research may appear to not pay off in dollar terms if we never live through a major disruption in the world's oil supply but that hardly means it has no value.

If we are going to have government sponsored research and development - why not do it through prizes rather than institutions? The current methodology of doling out large sums of money for institutions to achieve vague goals over many years, to see what turns up, is simply inefficient and guarantees waste. It is also not very transparent. Can you name the major goals of this years US research and development budget? For the money we are investing, we should all know them, and have a national debate about them.

Also I really laugh at the desperate attempts to create the meme that Government research created the shale gas play. Horizontal fracked wells are an obvious thing to try in low darcy rocks - you need to get h to compensate for the low k. And people were fracking to stimulate long before the Government got involved (the first frac was in 1948). The reason that the oil industry did not try horizontal wells earlier is that they are more expensive than vertical conventional wells in high darcy sands. The US was in a gas glut for many many years because there were so many cheap wells to drill. Why invest in shale in that environment? As soon as the price rose in the mid 2000's the industry took up the horizontal story and made it work by trying hundreds of different approaches in many different geological environments. The earlier work that was done by government subsided entities (and they count GRI as a Government entity!) was simply wasted, a few tens of tries was not enough to commercialize the technology, you need high prices and cost based on large scale operations to do that. The industry needed to develop the huge infrastructure that is needed to fracked the thousands of wells a year. This is not trivial - again there were many different business models tried. It is really amazing the power of private industry to innovate once the commercial prize is there. This trial and error process is hard on the individual investor (most fail), but great for society.

I somehow doubt that the 1948 fracking technology is the same technology being used today.

I think your view is pretty simplistic. Who paid for the fundamental research on the math (models of undergroud fluid movement, analysis of seismic data), physics (the thermodynamics, hydrodynamics, mechanics, materials science, geology, seismology etc), chemistry (interfacial phenomena, reaction kinetics, separation, research of chemicals like new surfactants and their properties etc).
Most if not all of this is government funded. The private sector only does the last few bits (application, development) and gets all the credit and profits. That' the truth.

Honza, In point of fact, frac design is still astonishingly primitive in terms of modelling how the fracs propagate in shale. Its mostly trial and error at the moment, based on learnings from the thousands of wells now drilled. But when the industry goes into a new area, often you are back at the start again. The models we do have use mainly some quite old and well known equations and insights dating back to Stokes, Reynolds, Poisson and others. Whether you consider someone researching in 19C Cambridge "government funded" or not, I would guess that is not very relevant to funding of research today. The models were generally developed by the oil companies, fine tuned using their results from thousands of wells. You may not be aware but most of the big oil companies have research arms, that do a lot of work in these areas. Being connected to the business, a lot more of what they produce gets used than for general non-Government based research where the decision as to what gets funded is based on a non-accountable bureaucrat.

The wider point is that you are like someone who used someone elses money to bet on a hundred horse races (while taking a cut off the top for services rendered). You then point to the 3 winners and say, "look my use of your money was a success". I think we all agree you have to include the losing bets in that analysis.

Actually, the entire system is quite transparent (well what they say they are going to do; many times they don't actually do what they say they are doing, but that's another story). You just don't know where to look. The abstract of every grant is available on a government site available online. All the RFAs are up there, too, for your perusal. So, you can certainly see where NSF and NIH sees as priorities for funding. Why we don't offer prizes? Well, because there's no class of people, unfortunately, who know better than the scientists themselves what is worth working on. Government types tend to be a decade or more behind the cutting edge. Unfortunately, the "success" of the HGP has empowered the government to take a more top down approach recently (e.g. ENCODE 2.0, "translational medicine," and now Obama's brain mapping boondoggle).

I don't have room to write an essay here, but from what I've seen during my years at the bench (alas, I'm in the private sector now against my will), the incentives for academic scientists are completely out of sync with what the public wants done. Because tenure is based on publishing C, N, or S papers which only publish novel, sexy stuff (which, in turn, is more likely to be wrong), scientists in academia aren't really doing what the public thinks they are paying them to do. On top of that, the push to get published in those top tier journals results in massive cheating. That's the dirty secret in academia. In fact, in my 20 years at the bench, I think I only met three people who I didn't see cheat out of the hundreds I worked along side.

This sort of happened with stealth bomber technology. The military set specific goals of what they wanted in an aircraft and whoever could provide the best option within those specifications were awarded with a military contract.

Government budget cuts should have two phases. Phase 1 should be the meat, the intended savings. Phase 2 should be a smaller amount, maybe 10x the average mgmt salary. Phase 2 is reserved for cost savings resulting from the firing of government employees who choose their cuts poorly.


There has not, nor will there ever be, any real foundational innovation emanating from the private sector. All innovation in the U.S. finds its source in DOD tech think tanks, like SRI, Aerospace Corp, Livermore labs and the like. Apple's innovation, contrary to the conventional view, didn't come from Steve Jobs, Steve Wozniak, Jonny Ive, or Xerox PARC. All of it, I MEAN ALL OF IT, came from SRI (Stanford Research Institute). If you want me to give names, well, Doug Englebart is a start (Mother of all demos, 1969. Pretty much everything in computing we've taken for granted in personal computing was seen in the culmination of a USAF sponsored research project, culminating in this demo.)

I'm not sure that your general statement is true, but to the extent that it is, it doesn't mean what I suspect you want it to mean. That a government-funded researcher was the first to develop a GUI doesn't mean that the GUI would never have been invented without government funding, or even that the privately-developed GUIs would have been significantly delayed if not for government funding. All it means is that the government had money to burn to research this stuff back before there was any real economic reason to be researching it. This isn't obviously an example of government doing things right, although I'll grant that it's certainly better than most of the other things government spends money on.

We would have never seen any of what we enjoy without gov funding of Howard Hughes aviation electronics hobby (tax shelter outfit) and the real secret behind silicon valley (SRI). There is no argument under the sun against this, because this is a self evident set of facts that are quite inconvenient for the ideologues.

Ideally you should read and attempt to understand a comment before replying to it.

Nah. No need to read really, but if you insists for a fair reading, well, OK. My fair reading, my fair fellow, Brandon, results in the conclusion that you ignore the absolute fact that the ball never gets rolling without a government nudge. The GUI thing would have never, ever happened. There were way too many capital and PHD intensive elements, all of a great variety, that the private sector could not have ever brought to bear in realizing things envisioned by visionaries. Only an ecosystem of think tanks and research facilities (Aerospace corp, SRI, Livermore) could pull what we've seen off.

Right, and the human eye is too complex to have evolved. so God must have created everything.

I halfway agree with you, but even so argument by assertion is a pretty poor way to convince anyone.

For those who contemplate any equivalence between gov R&D and DEA enforcement, well, their mere contemplation of this is a capital crime, infinitely worse than 9/11 crimes. There should be no questioning of Federal subsidy of R&D, and anyone who does should contemplate prison time. I'm actually serious on this. Freedom is not worth going into the 21st century, auguring back into the stone age. To those libertarian, rural Randy's: leave the planet. In fact, leave the universe, or any possible universe for that matter. Your no good for anything, including yourself, which should serve as a cautionary note as to your subscribing to libertarian nonsense.

All this crazy, mean stuff I say is simply emotional shorthand to express the fact that the private sector has neither the time nor the inclination for long term strategic technical thought, applying the skills and talents of a wide variety of Phd. level personnel focused in on very hard problems the private sector would never touch*.

*Note, Bell Labs was an exception, but that wasn't a real market oriented private institution, until the early 80's. The main constraint is that private capital, in the context of true market pressure, losses all patience required not only for the time required to realize solutions to hard problems, but to corral a wide variety of cutting edge talent, the focus of whom can only be directed to superordinate goals that transcend anything beyond the scope of traditional, private sector management ken.

kudo's to Tyler on article link, but it's old stuff, that should have been conventional established wisdom since the Lincoln admin (National Academy of Sciences and land grant universities). Alas, the ideologues are like nasty burnt fried egg residue stuck on the bottom of a nonstick pan--not amenable to that wonderful private sector technology, the brillo pad. I'm thinking, just throw away the pan.

"the ideologues are like nasty burnt fried egg residue..."

I agree. These people who would be OWSers, environmentalists, and pacifists in other contexts never fail to become cheerleaders for corporate welfare, fracking, and DoD technology spinoffs whenever it comes to discussing government R&D.

Let's not confuse inventor credit with innovation. Just because someone is given credit doesn't make them important. You can't say that we would be materially worse off without the inventor. It just means that somebody else would have received the credit.

Most spending, public or private, on medical research is, in the long run, deleterious. If someone has demonstrated, by falling ill or by involvement in an accident, that they are genetically inferior, why spend money trying to figure out how to keep them in the gene pool?

What is the "right" amount for the govt to be spending on this? How is that determined?

I am so glad you did not trace them back to Nazi's R&D projects.

The question with Gov. research funding is how much does it crowd out non-Gov. research.

As a recent PhD grad from a top 5 engineering college, I believe that academic R&D funding should be extensively cut and stream-lined for a more efficient production of new technologies. Having said that, the big question is how to go about accomplishing this...? I have a few suggestions personally; however, more discussion is need on this topic.

First and foremost, R&D is set up in a manner that creates incentives for professors/researchers/etc to publish papers in journals. The more papers you publish, the more grants you can write which leads to more funding. More papers also leads to getting tenure. The problem with this is that millions (or should I say billions) of dollars are used to produce papers that have no meaningful real life applications. From my experience, I would estimate this accounts for 85% of all research (and believe me, this is a modest estimation). I suggests that the incentives for funding should focus on having goal oriented funding (solving real world problems), and that the sustainability of the funding should be dependent on results of the research toward solving a problem (current funding sources have no feedback or repercussions once the funds are awarded). How to go about accomplishing this is anyone's guess.

Second, the theory that having more funding will create more researchers that produces even more technologies is false. A large portion of people that enter graduate school and become researchers, do so because they have limited other opportunities (research is not as lucrative as many might think). This is not always the case (there is a fraction that are passionate about academic R&D), but a large portion of people entering grad school do so because they cant find a job. That being said, these people do not get jobs because there are other better qualified candidates for these opportunities. So, by increasing public funding at academic institutions, you are getting a larger number of lower quality workers; thus decreasing the efficiency of the funded research.

These are just a few of my observations from performing both private and publicly funded research over the past 6 years. By all means, I am not an expert in publicly funded research; however, I believe it is a good topic for discussion (both the pros and the not so talked about cons of government funded research).

I'm not convinced either way, but I'd love to hear Tyler & company's reaction to terence kealey in that empirical evidence suggest that public funding of science crowds out privately funded innovation:\

It absolutely does! There is not much innovation going on in biotech industry. For the most part, all they do is refine ideas and developments that originate at academia. The main problem with this is that it is incredibly inefficient - all these bumbling, essentially self-taught graduate students doing the best they can while the boss is busy writing grants. Minimal or unexistent quality control, too. So a huge fraction of the output is a total BS. But there are so many of them that there is always enough decent output for industry to feed on.

Crowding out. Utter, complete nonsense. See the latest recent Atlantic article on unemployed/underemployed STEM PHD's. Also, one, just one, math or physics PHD who was forced to remain employed by taking a Wall Street gig due to a government research grant that was cut back is one PHD too many going into Wall Street. It's something akin to the Russian physicists who, because of needing to put cabbage on the table, takes that gig with the Iranian nuke agency. After all, what did WarrenB say about financial weapons of mass destruction.

$1 paid to employ a mathematician doing regular university mathematics research, keeping that mathematician from doing Wall Street work, has a payoff of $1 trillion. This is an absolute fact you can take to the bank, and is neither debatable nor arguable. So, fund all physicist and mathematicians enough money to keep them occupied and away from Wall Street, and you know what, the world just might be an ok place.

Wow, you have a seemingly inexhaustible supply of non-debatable facts. It seems to come in very handy. Has it occurred to you that funding all physicists and mathematicians will just encourage more people to go into math? It becomes a bit of a vicious cycle, and the effects of the marginal mathematician going into finance will eventually become much worse because the margin will be further out.

Comments for this post are closed