Fourth Franqui Lecture: Energy and the Industrial Revolution

The video of my fourth Francqui lecture on the energy and the industrial revolution is now on Youtube:

 

The opening graph of population and GDP per capita in the United Kingdom since 0CE combines data from the Maddison Project at the University of Groningen and data produced by Steven Broadberry. The energy data in the next graph was compiled in a 2007 publication by Paul Warde. The graph of energy use in Europe since 1500 and the graph of the composition of energy use are from "Power to the People" by Astrid Kander, Paolo Malanima, and Paul Warde.

The next section of the presentation gives a high level summary of Daron Acemoglu's theory of directed technical change and applies it to the two case studies. The first is my paper coauthored with Jack Pezzey and Yingying Lu, forthcoming in JAERE, on directed technical change and the British industrial revolution. The second is my 2012 paper coauthored with Astrid Kander on the role of energy in the industrial revolution and modern economic growth. As I mentioned in the lecture, we didn't know much about the theory of directed technical change when we wrote this paper and it didn't influence our research. Yet we can explain the results in terms of the theory.

The graphs that open the section on the British industrial revolution use data from Broadberry and Warde as well as from Robert Allen's book on the industrial revolution (the price data). The painting of the Iron Bridge is by William Williams.

Opening the section on Sweden is a photo of the Aitik copper mine. We used data from the Historical National Accounts of Sweden and Astrid's PhD research. If you are wondering how the value of energy could be as large as the GDP in 1800 in Sweden this is because energy is an intermediate good. GDP is value added by labor and capital with land included in capital usually. Gross output of the economy is much larger than the GDP. A huge amount of economic activity was dedicated to producing food, fuel, and fodder.

The solar panels that open the concluding section are in Japan. I've forgotten where.

Flying More Efficiently

I have another new working paper out, coauthored with Zsuzsanna Csereklyei on airline fleet fuel economy. Zsuzsanna worked as research fellow here at the Crawford School on my Australian Research Council funded DP16 project on energy efficiency and the rebound effect. This paper reports on some of our research in the project. We also looked at energy efficiency in electric power generation in the US.

The nice thing about this paper is that we have plane level data on the aircraft in service in 1267 airlines in 174 countries. This data is from the World Airliner Census from Flight Global. We then estimated the fuel economy of 143 aircraft types using a variety of data sources. We assumed that the plane would fly its stated range with the maximum number of passengers and use all its fuel capacity. This gives us litres of fuel per passenger kilometre. Of course, many flights are shorter or are not full, and so actual fuel consumption per passenger kilometre will vary a lot, but this gives us a technical metric which we can use to compare models.

The graph shows that the fuel economy of new aircraft has steadily improved over time. One of the reasons for the scatter around the trendline is that large aircraft with longer ranges tend to have better fuel economy than small aircraft:

This is also one of the reasons why fuel economy has improved over time. Still, adjusted for size, aircraft introduced in earlier decades had (statistically) significantly worse fuel economy than more recent models. We used these regressions to compute age and size adjusted measures of fuel economy, which we used in our main econometric analysis.

The main analysis assumes that airlines choose the level of fuel economy that minimizes costs given input prices and the type of flying that they do. There is a trade off here between doing an analysis with very wide scope and doing an analysis with only the most certain data. We decided to use as much of the technical aircraft data as we could, even though this meant using less certain and extrapolated data for some of the explanatory variables.

We have data on wages in airlines and on the real interest rates in each country. The wage data is very patchy and noisy and we extrapolated a lot of values from the observations we had in the same way that, for example, the Penn World Table extrapolates from surveys. There are no taxes on aircraft fuel for international travel and the price of fuel reported by Platts does not vary a lot around the world. But countries can tax fuel for domestic aviation. We could only find data on these specific taxes for a small number of countries in a single year. So, we used proxies, such as the price of road gasoline and oil rents, for this variable. We proxy the type of flying airlines do using the characteristics of their home countries.

The most robust results from the analysis – that hold whether we use crude fuel economy or fuel economy adjusted for size and age – are that – all things constant – larger airlines select planes with higher fuel economy, higher interest rates are associated with poorer fuel economy, higher fuel prices are associated with higher fuel economy (but the elasticity is small), and fuel economy is worse in Europe and Central Asia than other regions.

It seems that for a given model age and size, more fuel efficient planes cost more. This would explain why, even holding age and size factors constant, higher interest rates are correlated with worse fuel economy. Also, if larger airlines have more access to finance or a lower cost of capital they will be able to afford the more fuel efficient planes.

What effect could carbon prices have on fleet fuel economy? The most relevant elasticity is the response of unadjusted fuel economy to the price of fuel. This allows airlines to adjust the size and model age of planes in response to an increase in the price of fuel. We estimate that this elasticity is -0.09 to -0.13, which suggests the effect won't be very big. Because we use proxies for the price of fuel, we expect that the true value of this elasticity is actually higher. The elasticity also assumes that there is a given set of available aircraft models. Induced innovation might result in more efficient models being developed. There might also be changes in the types of airlines and flights. So the effect could be quite a bit larger in the long run.

Energy Leapfrogging (or Not)

Arthur van Benthem has a recent paper in the Journal of the Association of Environmental and Resource Economists titled "Energy Leapfrogging". The main thesis of the paper is that despite presumed improvements in the energy efficiency of individual technologies such as cars and refrigerators, energy intensity in developing countries today is similar to what it was in today's developed countries when they were at a similar income level. There is no "energy leapfrogging". This is also an implication of our paper "Energy and Economic Growth: the Stylized Facts". If there has been an almost constant log-linear relationship between energy use and GDP per capita then there is no energy leapfrogging.

van Benthem suggests that a major contributor to this is that the consumption bundle in developing countries today is much richer in energy services like personal transport than was the consumption bundle at a similar level of development in today's developed countries. Consumers have substituted towards these now cheaper energy services (what are they consuming less of though?).

On the face of it, this suggests that there would be a very large rebound effect due to substitution towards energy services. This is on top of any indirect rebound effect due to increased energy productivity boosting income and thus energy demand as originally proposed by Harry Saunders.

On the other hand, there must be some shift away from energy services as income increases so that energy intensity is lower in richer countries. Anyway, this is pretty speculative but worth thinking about, I think.

Economic Growth and the Transition from Traditional to Modern Energy in Sweden

During my recent visit to Sweden, we successfully revised and resubmitted a paper, titled: "Economic Growth and the Transition from Traditional to Modern Energy in Sweden". We have now made it available as a working paper in the CAMA Working Paper series.

This paper follows up on the paper we published last year in the Energy Journal. That paper looked at the paradox of energy and growth. How could energy have been important in the Industrial Revolution yet be a fairly small portion of production costs today. The new paper looks at the relative contributions of changes in the quantity, quality, and related technology (so called factor augmenting technical change) of traditional (biomass, animal power) and modern (fossil fuels and hydro-electricity) to economic growth in Sweden between 1850 and 1950. This was the period of the energy transition to modern energy in Sweden.

The graph shows that in 1850 less than 5% of energy in Sweden was derived from modern energy sources. By 1950 around 80% was. There are two large spikes in the share of traditional energy associated with the World Wars when imports of fossil fuels were restricted. The share of biomass in Sweden today is higher than it was in 1950. Because the share of modern energy was so small in the early years, even though the rate of improvement in the efficiency with which it was used was in fact higher than that of traditional energy, it contributed less to growth than traditional energy did. Over time, as the share of modern energy increased, this changed so that modern energy innovation and the increase in the use of (quality adjusted) modern energy contributed more to growth. However, according to our data and model, innovation in energy came to a halt towards the end of this period. By contrast, the role of labor augmenting technical change, which includes both better management of labor, increased human capital per worker etc. accelerated smoothly over time to become the most important driver of growth. Of course, Stern and Kander (2012) found this too. It's nice that the results in the two papers match! :)

The table presents the detailed growth accounting results. We actually computed these contributions for every year and then the table provides averages for each 20-year period. Betwen 1870 and 1910 growth was at first slower and then faster than our simple model fitted to the data predicts. But we found that giving the model more degrees of freedom to fit the wiggles in the data could lead to nonsensical results. It's also possible that it is the data that is mismeasured.

What the data implies is that it took a long time for the innovation in using modern energy to diffuse through the economy as the quantity of modern energy used increased. The following graph shows the rate of (factor-augmenting) technological change associated with each of the three inputs (the model also has capital but we assume its rate is zero):

We probably shouldn't take the negative values too seriously, but as noted above, fitting a more complex model was challenging. There was very rapid innovation in the use of modern energy in 1850-1890 starting at about a 7% per year increase in productivity and falling to about 3% a year. But the contribution to growth in the table started at 0.03% per year and rose to 0.08% per year over this time. Growth accounting type exercises, by attributing all the effects of an innovation to the year it happens, are extremely conservative. In later years, when the quantity of modern energy was much larger, that energy contributed more to growth than it would otherwise have done because those earlier innovations had permanently increased the marginal product of modern energy (ceteris paribus of course...).

Joel Mokyr is Very Optimistic on Future Technological Progress

I've made a few posts in the last half year or so on the potential slowing of the rate of technological change and economic growth and the implications of continued technological change for the distribution of income. Economic historian Joel Mokyr discusses some of these issues in a very optimistic piece on the future of technological change and economic growth. He sees no real limits to advances in technology and sees the implications for the future of labor as positive.

Capital Biased Technological Change

Another current debate is about the implications of capital biased technological change. It looks like there is going to be a series from Krugman on this. A previous blog by Krugman made the strange assumption that the capital stock was fixed. According to this blogpost, if you allow the capital stock to adjust workers are not made worse off in absolute terms by capital biased technological change though inequality rises assuming that some people mainly get labor income and some mainly capital income. Of course, it would make more sense to at least be using a constant elasticity of substitution production function with an elasticity of substitution that is different to one between capital and labor. You can't get biased technological change in a Cobb Douglas function without the ad hoc assumption that the elasticities change as the overall level of productivity goes up.

Results from Stern (2012) Energy Economics Now on Data Page

The results data from my recent paper in Energy Economics are now up on my data page.

Data is presented in terms of distances and log distances relative to a stochastic frontier estimated with the between estimator for the 1971-2007 period. A distance of one (or log of zero) implies that a country is just on this frontier. But the frontier moves over time as world best practice improves. Countries can have time series values below zero (or a distance of less than one) because they may reach levels of energy efficiency greater than the the average best practice for the whole period. This is especially the case in the later years. The file also gives income per capita in PPP terms from the Penn World Table. Click here for more posts explaining this project.

World Energy Outlook 2012 and the Rebound Effect

I have been reading the 2012 World Energy Outlook from the IEA.  There is a special focus section of three chapters on the role that energy efficiency improvements could play in reducing greenhouse gas emissions. The report is generally very conservative on estimated uptake of alternative energy and, therefore, efficiency will be needed if there is to be any chance of staying within a 2C trajectory.

There is, however, only one mention of the rebound effect in this whole section in Box 10.2 on p316. Somehow they come up with an estimated rebound effect of only 9%. This is almost certainly an underestimate of the rebound effect. Typical estimates for direct rebound in consumer applications are around 30%, while on the production side and at the macro-level rebound effects can be much larger than this. The report does correctly note that:

"A significant portion of this could avoided by appropriate pricing policy"

A cap on carbon emissions will induce energy efficiency improvements as part of the solution. Though there will still be a rebound effect it can't result in the emissions reduction goal not being met. However, an efficiency policy without a carbon cap is likely to yield disappointing results in my opinion. With a carbon tax, rebound means that the carbon tax would have to be higher than it would be if there was no rebound, I think.

Adele Morris' Presentation

Video of Adele Morris' presentation at the Crawford School has now been posted on our website. Please click on the link above to view it.

Energy Policy and Climate Mitigation in China: The Ideas Motivating Change

Based on reading of government documents and the writings of Chinese academics, Olivia Boyd's ANU masters thesis documents the role of three ideas in driving China's current energy and climate policies:

1. The idea of new energy security that stresses domestic, rather than international, sources of energy insecurity.
2. Green development and growing concern over the environmental and resource constraints on economic growth.
3. Low-carbon leadership, which posits a vision of China’s international political and economic influence based on climate leadership and low-carbon markets.

This is roughly what I have argued are China's motivations - for example in my CRWF 8000 lectures on energy and the environment that I gave in October - but I find a lot of resistance to accepting that China is serious about these issues and I based my view largely on conjecture rather than a close reading of the literature. This thesis stands on much more solid foundations.

I saw this paper on academia.edu, which I am finding more and more useful as people are posting interesting papers on it that I otherwise wouldn't see.

Some Kind of Deal is Agreed in Durban

So Kyoto will be extended but be restricted to Europe. In the meantime most other countries have their Copenhagen pledges to fulfill. Negotiations will proceed to agree a global treaty with "legal force" (whatever that means) by 2015 to come into force in 2020.

This is good, assuming things stay on track because we know that China needs to peak emissions by at least 2020.

I'm not too concerned about whether countries' commitments would limit climate change to 3.5C or 2C. The main thing is to continue the momentum that will foster the innovation that will solve this problem at a reasonable cost. We are seeing this technological change happening in a significant way.

Economics of the Singularity

I recently read part of this book: "The Lights in the Tunnel". I don't think it makes a lot of sense in terms of the outcome - the economy collapses from lack of demand as unemployment rises. But the basic idea is intriguing. Up till now machines and computers have taken over many tasks but they have acted as "q-complements" to labor in general and particularly skilled labor.* But could the future be one where all simple tasks are automated and the remaining jobs are too hard for people of ordinary ability to do?** And what happens when/if the "Singularity" occurs - when computers become more intelligent than humans? Does inequality continue to grow beyond anything we see today? Will redistribution increase? Is there a middle way? Is this whole idea totally mistaken? Or will humans be enhanced by genetic manipulation or implanting of non-biological computers?

There are economists and entrepreneurs/technologists who have thought about this. But I don't think anyone has any real answers yet.

* If an increase in an input increases the marginal product of another input then those two inputs are q-complements.
** More precisely the most that anyone is willing to pay for a human to do these automatable tasks is less than the minimum amount needed for subsistence.

Special Issue of Energy Economics

The latest issue of Energy Economics is a special issue on the economics of technologies to combat global warming. I haven't read the papers in detail yet, but there look to be several good overviews of the current state of play on the key issues in this area. Given the slowness with which countries are moving to put a price on greenhouse gas emissions, technology policies might end up being more important in driving the transition to a lower carbon future.

The special issue is edited by William Nordhaus and Nebojsa Nakicenovic. I'm going to be reading them in preparation for the lectures I will give later this year in the Crawford course Government, markets and global change. My lectures are on the general topic of energy and the global environment with a particular focus on alternative energy technology policy. The course is a team taught course that uses cases studies to structure the lecture content. We are using a case study on US DOE funding for alternative energy technologies. I imagine these papers would also feed into writing my sections of our IPCC 5th Assessment Report chapter.

Why "Jevons Paradox" is an Argument for Stronger Action on Climate Change not Weaker

The 19th century economist William Jevons suggested that improvements in the efficiency of energy use devices could increase energy use rather than reduce it. The more efficient machines etc. would reduce the cost of production thereby increasing the amount demanded and sold and, therefore, the energy used to produce the products. The rebound effect is a modern statement of this idea: "Efficiency improvements will reduce energy use by less than the efficiency improvement". In the case of consumers, the rationale is that efficiency improvements are equivalent to reductions in the price of "energy services" such as heating, lighting, air-conditioning etc. The law of demand tells us that this will increase the demand for these services and, therefore, for the energy used to produce the services. The size of this rebound effect is an empirical question.

There have been some articles in the popular media about Jevons' paradox recently. The Economist argued that we would be better off without lighting efficiency improvements because they will result in increased energy use and hence pollution. Roger Pielke argues that Jevons paradox tells us that we both need to increase energy efficiency and energy supply in the future. On the other hand, the Climate Progress blog tries to debunk the Jevons' paradox while admitting that the rebound effect is real.

Instead, I argue that the rebound/Jevons' effect tells us that the results of direct action on climate change are likely to be disappointing. Efficiency improvements would need to be bigger than the desired savings in energy use. But, by contrast, a cap on carbon use would eliminate the "carbon rebound effect" due to efficiency improvements. We really do need improvements in energy efficiency as part of the solution to climate change and they make us better off. But a carbon price can be part of a more effective policy.

Artificial Photosynthesis

Interesting innovation - solar furnace that uses sunlight to convert carbon dioxide and water to carbon monoxide and hydrogen. These are necessary inputs in the Fischer-Tropsch process of synthesizing liquid hydrocarbons. So, this is effectively artificial photosynthesis. At the moment, the efficiency is no better than plant-based photosynthesis and the reactors would be more expensive than harvesting crops presumably. But with efficiency gains this might be a viable way to generate liquid fuels for aircraft and other applications where current electric technologies are not viable. Seems that it is one more nail in the coffin of the "hydrogen economy". Separating the low density hydrogen here would make no sense, I think. An article in Science provides details.

Porsche Develops Hybrid Technology

I blogged about the development of hybrid cars by BMW and Mercedes when I was visiting Munich. My point was that fuel economy standards were forcing luxury car makes to adopt hybrid technology. I saw this as a route to wider adoption of hybrid technology in mass-market cars. Non-luxury hybrids seem so far to only appeal to "green consumers" willing to pay a premium for lower fuel consumption. Anyway, the New York Times has an article on new hybrid systems developed by Porsche.

One is a hybrid version of the Cayenne SUV with a 35kW electric motor and a 250 kW petrol engine. Urban fuel economy is improved from 16 mpg to 21 mpg. Highway fuel economy only improves by 10%. Then there is a racing car that uses a flywheel to store energy from braking which can then be used to power a generator. It has two 60 kW electric motors driving the front wheels in addition to the 360 kW petrol engine. The third system is a concept car that has a 375 kW petrol engine and front and rear electric motors that can produce a total of 164 kW. For comparison, the current standard V6 Ford Falcon has a 195 kW engine.

There is an increasing diversity of body plans out there, which is a positive sign in the development of a new technology.

Tsao & Waide: "The World’s Appetite for Light"

Following up yesterday's post I looked at the study that provided the background for the paper studied there:

Jeffrey Y. Tsao and Paul Waide:
"The World’s Appetite for Light: Empirical Data and Trends Spanning Three Centuries and Six Continents"
LEUKOS VOL 6 NO 4 APRIL 2010 PAGES 259 – 281

This paper gives a figure of per capita light consumption and GDP/cost of light:



I drew the thick black line on the chart which represents an alternative curve fit. I know it is very crude but I couldn't draw a nice smooth curve with my software. I'm not saying that this is a best fit curve, just that it is a vaguely plausible alternative. If something like this model was fitted then the rebound effect is less than 100%. Tsao and Waide do not test alternatives to their linear model that assumes an income elasticity of one and a price elasticity of minus one of lighting demand. I think more research in this area is definitely warranted though I praise Tsao and Waide's pioneering attempt to bring together different sources of data and begin the empirical analysis.

Whatever the truth, the Economist's proposal to just stick with incandescent lighting is wrong. Even if there were no energy savings from introducing solid state lighting, as Tsao et al. state people would have more lighting services for a given energy input. There are environmental impacts to having too much outdoor light. These should be addressed separately, not by halting technological progress. If regulation limited the amount of outdoor light then these innovations would result in saving of energy...

Will the Adoption of Solid State Lighting Lead to an Increase in Energy Use?

The Economist discusses an article in Journal of Physics D: Applied Physics by Tsao et al. on the effects on energy use of the adoption of solid state (i.e. LED) lighting (SSL) on global energy use. The Economist argues that it would be better to keep incandescent bulbs as a result. This seems a bit crazy. The literature on the rebound effect suggests that for energy saving innovations for consumers in developing countries the rebound effect is typically of the order of 30%. In other words, the net energy savings are around 70% of the amount of energy nominally saved by the innovation. Joshua Gans comments on this article taking a direction inspired by the Schumpetarian endogeneous growth literature where new innovations are sold by monopolist innovators.

To understand why the authors posit such a large rebound effect I took a look at the original article. The first key leg of their model is the following relationship between historical data on lighting use and a very simple model:



Most of the early data relies on the work of Fouquet and Pearson. The model treats light consumption as a function of two variables: GDP and cost of lighting. The elasticity of demand with respect to the cost of lighting is minus one and the income elasticity is plus one. Based on this data the model looks pretty plausible. It would be nice though to see this data in per capita terms or to see lighting intensity of GDP plotted against cost of lighting to get a better idea of how robust it is. The way in which the cost of lighting is computed will be very critical too. I'll look at that in a subsequent blogpost.

If the demand elasticity is minus one then any reductions in the cost of lighting will be exactly offset by increases in consumption of lighting services. Both these elasticities seem high in absolute value for developed economies. So it would be nice to at least test a model which allows the elasticities to vary with income level vs. a model which does not if you are going to make big predictions about the future.

Solid state lighting will certainly reduce energy costs of lighting but this is achieved partly by substituting capital for energy. At the moment, LED lights are expensive. This means that the cost of lighting is reduced by less than the energy use is reduced by the innovation currently. Therefore, even if the price elasticity of demand was minus one, adoption of solid state lighting would reduce energy use (ignoring indirect energy costs of capital). The authors argue of course that these costs will reduce rapidly. Historically, they argue that capital costs are typically 1/3 of energy costs of lighting. They assume that by 2030 the capital costs of SSL will be the same so that there is no capital-energy substitution in the adoption of SSL.

Role of Fuel Economy Standards in Effecting Technological Change

I was fascinated to learn yesterday that Mercedes has paid the US government $300 million in fines due to violating the CAFE standards. As fuel economy standards continue to rise manufacturers will need to find innovative ways to reduce the fuel consumption of their luxury and sports models in order to end up with acceptable cross-fleet mean fuel consumption figures. The interesting thing is that luxury cars is exactly where it is easiest to absorb the costs of new technologies. Most automotive innovations have been introduced first on luxury models and then trickled down to mass-market vehicles. Why should fuel economy technologies be any different? But up till recently, manufacturers have tried to sell fuel efficient mass-market cars. Some of the more innovative such as the Prius have only been competitive beyond a green signalling niche market with government subsidies. But experimenting first with up-market models makes more sense to me. There are already Lexus hybrids but now BMW and Mercedes seem to be following suit.

Today, I went to the BMW Welt museum/exhibition here in München where I happen to be visiting:



A major theme of the exhibition was new technologies to reduce fuel consumption while maintaining performance. The suite of technologies is termed BMW Efficient Dynamics. Several BMW models already incorporate the technologies, which include regenerative braking, engines which shut off when the car isn't moving, more efficient fuel injection etc. Hybrid vehicles will be available soon.

BMW is still pushing hydrogen cars, which seem to be a mistake to me. Especially, using hydrogen to fuel an internal combustion engine makes no sense at all I think.

The Environment and Directed Technical Change: Acemoglu et al.

Acemoglu, Aghion, Bursztyn, and Hemous put out an interesting NBER Working Paper last October. The abstract is below. They carry out a simulation which shows that a carbon tax alone is significantly inferior in terms of loss of consumption to a combination of a carbon tax and clean technology development subsidy. Results depend on the elasticity of substitution between dirty and clean inputs and the discount rate. If the elasticity of substitution is 10 then temperature never rises by more than 1.76C irrespective of the discount rate and a carbon tax only policy costs 0.92 to 1.55% of consumption relative to the optimal policy depending on the discount rate. But lower elasticities of substitution (5 or 3) make a carbon tax worse (2-4% consumption loss) and result in catastrophic climate change (7-8C) under higher discount rates (1 to 1.5% rate of time preference, latter is Nordhaus' choice).

It is likely that the interfuel elasticity of substitution is greater than unity. But, based on my research I think it is very unlikely to be as high as 5 or 10.

Based on this research more attention should be paid to combining innovation policy with a carbon tax. This is a position that is, I believe, advocated by Roger Pielke among others. But it also shows that there can be a huge difference between using discount rates as high as 0.015% rather than the 0.001% favored by Nicholas Stern in assessing climate policy.

There is a lot more besides this in the paper including the effects of delay and non-renewable resources and the problem of global policy coordination,

Abstract
This paper introduces endogenous and directed technical change in a growth model with environmental constraints and limited resources. A unique final good is produced by combining inputs from two sectors. One of these sectors uses "dirty" machines and thus creates environmental degradation. Research can be directed to improving the technology of machines in either sector. We characterize dynamic tax policies that achieve sustainable growth or maximize intertemporal welfare, as a function of the degree of substitutability between clean and dirty inputs, environmental and resource stocks, and cross-country technological spillovers. We show that: (i) in the case where the inputs are sufficiently substitutable, sustainable long-run growth can be achieved with temporary taxation of dirty innovation and production; (ii) optimal policy involves both "carbon taxes" and research subsidies, so that excessive use of carbon taxes is avoided; (iii) delay in intervention is costly: the sooner and the stronger is the policy response, the shorter is the slow growth transition phase; (iv) the use of an exhaustible resource in dirty input production helps the switch to clean innovation under laissez-faire when the two inputs are substitutes. Under reasonable parameter values (corresponding to those used in existing models with exogenous technology) and with sufficient substitutability between inputs, it is optimal to redirect technical change towards clean technologies immediately and optimal environmental regulation need not reduce long-run growth. We also show that in a two-country extension, even though optimal environmental policy involves global policy coordination, when the two inputs are sufficiently substitutable environmental regulation only in the North may be sufficient to avoid a global disaster.