Innovation Files has moved! For ITIF's quick takes, quips, and commentary on the latest in tech policy, go to

To Deploy or Not to Deploy?

A recent report by the California Council on Science and Technology has rekindled the debate (see Andrew Revkin, Joe Romm, Dave Roberts) over technological readiness in clean energy, and whether we should be committing resources to innovation or deployment. I’m going to argue here that this deployment question is something of a false choice: the question isn’t whether to deploy or not to deploy, the question is how. It’s really about balance: getting this balance right means we optimize scarce resources to give ourselves the best shot at slowing emissions; getting it wrong means wasting time and money, chasing false solutions while emissions grow.

First things first: the California report asked whether existing technology can achieve the state’s aggressive emissions targets, and the results were straightforward: looking at technical feasibility, and assuming neither political will nor economic cost constrains the effort, California could hypothetically reduce its emissions to 60% below 1990 levels by 2050 – a big number, but short of its mandated target of 80% below 1990 levels. Meeting this more ambitious target requires new technology:

All of the approaches that will reduce emissions from 60% below 1990 levels to the target value of 80% below 1990 levels are going to require significant levels of research, technology development, invention and innovation. This part of the problem is therefore as much a technology problem as it is a policy problem.

Few would be surprised by this finding: that we need better, more affordable clean technologies if we’re to ever realize a full transition away from fossil fuels. Better car batteries, scalable biofuels, low-cost CCS and solar – we need it, and no one quite has it yet. Most can agree on this general point.

The real question plaguing the debate, rather, is where to get this better technology. Yes, most would say “innovation.” But “innovation” in itself can mean very different things. It’s useful to think of innovation on a rough (and admittedly simplified) sliding scale: on one end you have “incremental” innovation, and on the other end you have “radical” innovation. When a car company steadily modifies its engine designs for efficiency, or adopts new processes on the factory floor, this tends to be “incremental.” It’s experience-based and doesn’t require fundamentally new knowledge. But when that same car company begins installing IT-based automation in its vehicles, or pursues new battery chemistries for electric cars – we slide towards the “radical” realm.

An important point: innovation at the “radical” end is much less attractive for private companies or entrepreneurs in most industries to pursue on their own. Here’s where government steps in to assist or outright do it all (which is why the federal government funds so much basic research). For many in the debate, like Romm, deployment – meaning subsidies, standards, carbon prices, and other such tools – is the primary means of generating innovation. But deployment policy sets up existing firms as the main innovation drivers, and thus is suited to deliver the more-incremental, less-risky innovations that businesses in most industries are willing and able to do.

If incremental-type innovation was all we needed, then deployment would be fine; but if we take this rough framework and apply it to the current status of clean energy technology, it’s clear that incremental innovation has an important role to play, but in large part will not cut it. To paraphrase Schumpeter, you can build as many ethanol distilleries as you like, and you will not get algae biofuel.

It’s true, incremental innovation can help drive new business models, address integration, and resolve other challenges in areas like solar. But it won’t create a step-change up to next-generation nano-based solar modules, when the current generation of panels hit their performance/cost barriers. Nuclear power plays a big role in the California scenario, but deploying more Generation III reactors won’t cause spent fuel recycling technology to spring forth. There are major limits to existing battery and biofuel technology as well, that won’t be solved by putting more electric vehicles on the road or ramping up production mandates. All of these challenges ultimately require the “radical” kind of innovation. Deployment to drive incremental innovation could have some benefits in some areas, and we should be cognizant of those benefits, but the mileage on these will vary. We’ll have more on this point in future posts.

I’d argue it’s a different story for efficiency and smart grid, where the case is far stronger to pursue it aggressively now. It’s also a different story for wind power, where the need for “radical” style innovation is smaller. This means there’s in fact a case for limited, targeted, innovation-oriented deployment. Such a strategy might include a moderate clean energy standard that recognizes regional resource differences and is structured to reward performance gains or penalize power producers that can’t reduce costs at a reasonable rate. A limited reverse auction would achieve a similar goal, enabling incremental learning without wasting more subsidies than are merited. Indeed, if done right, deployment policy becomes a necessary subset of innovation policy, as it should be.

All of this requires being conscious of where technologies stand in relation to their dirty competitors nationally, what the future might hold in emerging alternatives, and a general awareness of what kinds of innovation are needed. These are not questions that often get asked, but they need to be.

But we also need to understand that, even if deployment has some benefits, there are tradeoffs. We are operating with a limited resource pool, especially in these times. A billion dollars we spend on a production subsidy to prop up low-upside technology with ultimate limitations is a billion less we have to spend on the next generation of high-upside technology. Rather than throw money (and political capital) into subsidies haphazardly, we need to be vigilant in their application, lest we end up with ethanol credit version 2.0. Imagine if, instead of pumping money into ethanol production, we had used all or most of that money for research on advanced biofuels that can actually scale. Where would we be now?

Lastly, we have to remember that what’s possible for deployment in the US will not be possible everywhere. Report co-author and Livermore Lab principal associate director Jane Long, writing in Nature (sub required), outlines the scale of the effort:

[California] could reduce emissions a lot — by perhaps 60% below 1990 levels. But it would have to replace or retrofit every building to very high efficiency standards. Electricity would have to replace natural gas for home and commercial heating. All buses and trains, virtually all cars, and some trucks would be electric or hybrid. And the state’s entire electricity-generation capacity would have to be doubled, while simultaneously being replaced with emissions-free generation. Low-emissions fuels would have to be made from California’s waste biomass plus some fuel crops grown on marginal lands without irrigation or fertilizer.

Let’s be clear: this is a lot to ask for, requiring the equivalent of a wartime-level mobilization, over the course of decades. Brad Plumer says that achieving such aggressive goals would be “mind-boggling.”

So if the report is right, existing technology can only solve three-quarters of the problem even if ideal political and economic circumstances emerge to facilitate deployment, circumstances that do not yet exist and may well never exist. This is not to say that these measures are uniformly impossible, nor is it to say that we shouldn’t fight for many of them; it is to say that they’re really hard, incurring major costs and inviting many political battles. What’s possible and what’s achievable are very different things. The less success you have with this deployment package, the more you’ll need radical innovation to make the transition more palatable – which, of course, you need to do anyway.

If we’re serious about a deployment strategy, we’d have to believe that all of the battles described above could be resoundingly won and then duplicated everywhere – not just in California, but in West Virginia, and South Africa, and India, and Spain, and China. Let’s say, for argument’s sake, every single one of these cogs falls into place in California: the public is willing to subsidize ever-growing amounts of clean energy, and embraces nuclear power with open arms, and is willing to buy more expensive electric vehicles. Will the same be true in every society everywhere else in the world? Of course not.

We are much better off optimizing our innovation system to deliver technologies that can be sold wherever there are burgeoning populations attempting to climb out of energy poverty. Incremental innovation, through deployment, no doubt has a role to play. But if deployment is our lead strategy, we never get where we need to go.

Print Friendly, PDF & Email

  • alannogee

    I agree with a lot here. It’s very helpful to narrow differences. Particularly with respect to nuclear (and CCS), breakthroughs are needed, and more deployment of current generation is not going to significantly affect costs. The ability to achieve economies of scale simply haven’t been demonstrated for these techs; nuclear has always been going the other way–getting more expensive over time. Breakthroughs with small modular reactors might be able to reverse that trend, but it is too soon to tell.Also appreciate the agreement that techs that are cost-effective or close to it, like many efficiency techs, and wind are ready to scale through deployment policies. For solar, batteries, and other more expensive techs, I don’t agree that radical breakthroughs are necessarily critical, though they would clearly be enormously helpful. I would argue that we can’t know whether R&D breakthroughs or incremental industry learning is more likely to get the job done, but that the latter is perhaps equally likely, and will accomplish a lot in any case. R&D breakthroughs are somewhat more of all or nothing propositions. You can bet the farm and lose. So while we need both approaches, I think deployment is still the lead strategy.Case in point. When DOE massively scaled up R&D during the Carter years, they focused on breakthroughs with 1 MW and 5 MW wind turbine designs that were much bigger than then-existing designs. They utterly failed to develop workable products. At the first national renewable energy conference I attended, around 1995, AWEA’s director said that by incremental scale=up, the industry had achieved reliable, “optimal-size” turbines — 200kW. Today, of course, through continued incremental learning, the industry standard is 2 MW with 7 MW and bigger machines on the horizon. Generating costs fell by more than 80 percent and wind became competitive with new fossil generation.So yes, the necessary political conditions for deployment are hard and may never exist. But the radical tech breakthroughs in the lab, and in the jump from lab to field, may not happen either. We therefore need both approaches to hedge against failure of the other, and it’s likely that both will make critically important contributions. But we still have to solve the deployment problems for R&D to succeed, and we can make great progress toward the goal even if radical R&D breakthroughs do not occur. The same cannot be said, however, if we bet too much on R&D. Finally, as per my comments to the Revkin post, the same percent carbon reductions are actually easier in other regions and countries that are more coal dependent. And the Long study, in my opinion, undervalues the role that high penetration of electric vehicles and heat could play in providing storage for a high-renewables system, even with incremental progress in existing technology. Alan Nogee,Clean Energy Consultingformer Director, Union of Concerned Scientists Clean Energy Program

  • Matt Hourihan

    Hi Alan – A couple quick responses, starting with nodding agreement that ultimately we’ll need both approaches and that both are critical.It seems that much of the disagreement here comes down to what’s possible versus what’s not, in the realm of incremental learning as it relates to specific technology areas. I’m very familiar with the wind story, and see the experience in places like Denmark as a good example of what can happen when technology-push and demand-pull are working together in rough synchronicity (and see the US wind story as a good example of what happens when they’re not). But then, the challenges of wind power seem roughly more suited to low-risk, incremental gains and engineering fixes anyway. With solar, on the other hand, the story is slightly different: while the slow and steady gains might be useful in, say, balance-of-system cost reductions, the contributions of “radical” innovation seem much more relevant for other pieces of the puzzle like module efficiency, where applied science very much has a major role to play still. I don’t know how you get scalable, affordable solar in as little time as possible without a strategy that addresses all of these elements, and no one else does either in spite of claims to the contrary.This is as much about recognizing the limits of “deploy deploy deploy” as it is about recognizing the long-term upside of R&D investment, which is substantial if history is any guide. See, for instance, how SunShot is going about the challenge, attacking at multiple levels in an interdisciplinary fashion, inside and outside the lab. And I’m not sure I understand your concerns about betting too much on R&D, because we are a LONG way from making that “mistake.”I’d make the same argument for biofuels and batteries: no one actually knows how to make this stuff work at a level on par with fossil fuel costs and performance, if all we can do is try to improve what we already know. We can continue growing the market for electric cars in a limited and cost-effective fashion with an eye towards driving that incremental innovation, but at some point those improvements will flatten, or the cost of propping up the market will become too unwieldy. And even in areas like efficiency or grid, there are no doubt some big leaps forward that could be achieved through technology-push, in areas like power electronics, so again we need to be wary about doing too little as a country on these more radical challenges.And again, deployment has a role to play so long as you’re a) making these cost declines an explicit policy goal and b) not betting the farm on it, either, at the expense of also doing technology-push. This is not (and we are not) anti-deployment, but if we are critical of deployment, it’s because deployment is frequently oversold, and NEVER has innovation been a central concern for most in the energy and climate community at any rate.