Is it better to invest in subsidies, or in R&D

If you want to improve renewable energy, or access to health care, is it better to invest in R&D or to invest in subsidies?

The goal of subsidies is to make something affordable so that people can either afford it in the first place, or so that it becomes more affordable than more polluting or less desirable alternatives, in the goal that once that happens market dynamics will cause people to pick the better option.

However isn’t R&D the better investment? For example, if we spent $50 billion a year on health care subsidies/tax credits, wouldn’t that money be better spent on $50 billion a year researching and developing ways to provide higher quality care for less money so people could afford health care w/o the subsidies?

With electric cars, is it better to invest $5 billion into R&D to increase battery lifespan and decrease costs, or is it better to invest that money in subsides?

If you invest in subsidies, you seem like you are starting the market up which will drive down costs by mass production and more money for innovation which will increase reliability and decrease costs. But if you invest in R&D you do the same thing (eventually) because the lower cost and higher reliability will result in mass production anyway.

Is it better to invest $10 billion in solar energy subsidies or $10 billion into solar R&D to lower cost of production?

Is there a meaningful difference in price, quality, job creation, adoption of new technologies, universal access, etc. if you invest in R&D vs. subsidies?

This is an easy one - the answer is subsidies.

If you put money into R&D, you have to pick one or several projects to fund. If 10% of your money goes to winners you are doing very good. That is just the way R&D on any significantly challenging project goes.

Subsidies, however, encourage a market that is nascent. Perhaps subsidies for solar cells go to ones that aren’t as efficient and are more expensive than they could be. However, the subsidies have some positive impact guaranteed. Even better, they will encourage a lot of potential businesses to invest in R&D to get a chunk of this new market, to the level that will probably exceed what the government would pay for, and in a way that is more flexible than what a government run program can be. The same 90% will fail, but that will be paid by VCs and not the government. Plus. that crazy idea which gets rejected by the review board might be the one to solve the problem.

There is a lot of silicon valley money and talent going into solar cells, inspired by current and expected subsidies for them. Hybrids took off in California partly because of the subsidy of letting them into the car pool lane. R&D is good for way far out there stuff where subsidies make no sense.

it’s better to invest in prizes. Distribution of grants to researchers is likely to be politicized in various ways, but a prize for accomplishing something easy to measure and obviously useful (maybe a solar panel that sells for $X per square feet, generates average power so-and-so and lasts so long) - that is going to encourage getting things done, not politicking.

Subsidies, as Voyager says. The government trying to pick a direction for possible avenues of research is, well, not got a very good track record. A bit of thought will tell you why, as bureaucrats aren’t exactly noted for their foresight. Also, there is the vote buying and populist aspects, as well as the whole empire building thingy…bad bad.

One of my professors once gave a lecture that was talking about this very subject. To make a long 3 hour lecture short, he was talking about What If the British government, at the height of their power (say just after the Napoleonic wars) had decided they wanted to fund research and development in wireless communications in order to allow for better communications for the empire. What would they pour their research into?

Going through the permutations, at least according to my professor, the least likely thing they would have stumbled on was the findings of a little known guy named Maxwell, which would mean that, while there might very well have been some interesting spin-offs, they probably wouldn’t have had a hope in hell of directing research into the channels that would have actually produced something like a radio.


How does this square with the notion that tech tends to accelerate during wartime? Stuff like radar, the atom bomb, jet aircraft, etc. were all products of government R&D during WW2. It is hard to contemplate a project more bureaucratic and goal-oriented than the Manhattan Project - and it did indeed work in creating a wholly new tech - nukes.

Perfectly IMHO. Radar, the atom bomb, jet engines, they were all developed BEFORE the war (or at least the theories and principals were). Certainly, during the war the various governments were able to pour money into projects to develop those technologies, but they weren’t really researching them, more like engineering already conceptualized technologies and finding military applications for them.


I think we’d need to look at how many discoveries are made by industry as opposed to academia/public grants.

I believe that generally it ends up that most new discoveries come from academia, while as industry simply refines and optimizes things to get them smaller, faster, and cheaper, but without making any particular new discoveries (beyond the technology used to improve the technology.)

If you’re happy enough with wind, solar, nuclear, etc. then optimizing is all you need–if there’s theoretically still room to be made. If you want the next generation of technology, then you’re not going to get it out of industry.

By the same token, then, could not R&D be used to make practical theories and principles merely laboratory-imagined?

The atom bomb was only conceived of in principle prior to the Manhattan Project. No-one knew if it would work in any practicable form. If government R&D can be used to make such principles into practical realities, is it not useful?

I think investing in improving supply makes sense. Unfortunately, certain politicians seem to think that “the market” will “magically” do this if we just cut taxes enough, & others seem content with demand-side spending.

In the private sector, the company that does the research can create long-term efficiencies–if they get it right. But they have to sink the R&D money in; & a lot of businesses are timid; or perhaps they are on uncertain margins, & can’t afford the R&D; & so they would rather coast.

The government can drag the country forward. It can fund R&D for the sake of the whole country. But that would require some actual sense of what makes the most successful businesses in the private sector succeed, & it’s just too easy to sneer at government research.

And…that’s not what you asked.

Yeah, I considered that a form of R&D. If the federal government gave out 10 billion dollars for hundreds of prizes for advance X in field Y, then it would spur innovation and hopefully bring products to market faster.

I get the impression that subsides really wouldn’t make a difference early in a technology, but once the technology is advanced enough that it is nearly on par with existing technologies, subsidies would help.

I don’t see how subsidies would help wind power in 1980 when it cost 40 cents a kwh (vs about 5 for coal). But once wind hit about 7-8 cents a kwh, then subsidies would make a difference.

So that is my impression, R&D is best when you are far away from market parity, then subisides when market parity is obtainable with affordable subsidies.

I don’t know. I was thinking about things like Medicare, where we subsidize health care on the demand side more than the supply side, & prices–well, they don’t go down…

I would lean toward yes, subsidies, but I’m not sure what the mix of vouchers vs. direct funding of the installation should be.

Two things. In that case, clearly there was no way of doing either subsidies or involving private industry. Second, remember the Germans didn’t do such a good job with their government funded R&D. Anyhow, in that case we knew the answer, and were able to fund several alternatives. The way of cutting costs for solar cells isn’t as clear.

True, but any given company doesn’t do any better. I worked for Bell Labs, trust me, I know. When you have 10 companies working on something, 1 wins and 9 go bust, so it might seem industry does a better job.

Things like NSF and NIH, which give small amounts of money to lots of investigators, do better though most of those projects don’t do much either.

Certainly. The devil in the details can be summed up in one phrase, however: which ones? Which technologies should the government focus it’s (our) R&D money one? There are plenty of theories out there, plenty of theoretical scientists…which of the myriad out there should the government focus a Manhattan Project level of effort on?

I don’t believe this is the case. I think they KNEW it would be ‘practical’ (i.e. it would blow the crap out of stuff)…the thing was, they weren’t sure what the engineering would encompass. The US got into the game basically because a group of scientists convinced the president that there was a good chance that the Germans were already working on the thing.

To answer your last question, I don’t think anyone denies that government funding is necessary. Some projects are pure theoretical science, which industry is not going to want to fund heavily as there is no guarantees that any one avenue of research will pan out…or that, even if it does pan out it won’t take years or decades to do so. But when the government attempts to DIRECT R&D efforts, that’s where it runs into trouble, because they don’t have a crystal ball either, yet they have tremendous resources and all the drawbacks of any bureaucratic system that has it’s strings being pulled by politicians who are at least nominally beholden to a fickle public. And this leaves aside the corruption aspects that are inherent in anything like this.

That’s why subsidies are better than directed R&D, at least IMHO.


I think government spending on R&D is appropriate when the market is such that it is unlikely the goal can be accomplished through market incentives.

For example, let’s say your goal is to reduce power consumption by appliances. Well, you can implement taxes or subsidies (I’m lumping them together for this discussion) that incentivize the appliance market towards lower power consumption. Because the market is big, because there tends to be relatively high turn over for a lot of appliances, and because there are a number of players in the market, you’ll probably be able to achieve the goal this way. The benefit here is that multiple players will be trying to achieve the goal in different ways, so hopefully you get multiple ways to achieve the goal.

Now, let’s take another market that’s related - power distribution. There aren’t that many players in the market, the market isn’t really a competitive market (generation might be competitive, but I wouldn’t say that distribution is), and there are high infrastructure costs, so the technical turnover can be quite low. If your goal is to improve power distribution efficiency, you probably want to fund direct R&D (and supplement that with the appropriate tax or subsidy structure).

That wasn’t really the point I was making there. And I worked for Bell Labs for a while in the 80’s myself, so I’m aware that companies don’t necessarily do better at directed R&D either.


I think you have to have a balance. Don’t put all your money into subsidies or R&D. Invest some in both. I would think that 90% subsidies and 10% R&D would be a good mix, but that is just my wild guess. You might also want to put a little money into educating people to exactly what the problem is and how it will impact them.