The demand for energy and climate policy
Conny Olovsson
November 2012
Main question: how should climate policy be designed?
We need to understand the demand and supply for fossil energy to answer this question.
We also need to understand the evolution of technical progress.
Is it going to help us if/when we must reduce fossil fuel consumption for climate reasons and/or limited supply?
Policy recommendations must be based on models that are consistent with the real world.
How should energy be modeled at the macro level?
Almost all of the past macroeconomic research has abstracted from energy as an input into production.
Its share is (very) small: around 5 percent.
It might not be quantitatively important for understanding macroeconomic phenomena?
Energy is the link between the economy and the climate so it must enter the analysis, but how?
the way energy enters the production function is crucial for the policy conclusions we arrive at.
Previous research
A large literature on natural resource economics was triggered by the oil price shocks in the 1970s: Stiglitz, 1974, 1980; Solow, 1974;
Dasgupta and Heal, 1974.
Theoretical, not quantitative.
Abstracted from (endogenous) technological progress. Showed that with Cobb-Douglas production function, Yt =AtL1 α γKtαEtγ,
balanced growth with constant income shares is possible even if E comes from a source in …xed supply (e.g., oil);
oil use should be (exponentially) decreasing over time;
and oil use is independent of technology, population growth, etc.!
The Cobb-Douglas framework still used by Nordhaus and Stern.
Previous research
A large literature on natural resource economics was triggered by the oil price shocks in the 1970s: Stiglitz, 1974, 1980; Solow, 1974;
Dasgupta and Heal, 1974.
Theoretical, not quantitative.
Abstracted from (endogenous) technological progress. Showed that with Cobb-Douglas production function, Yt =AtL1 α γKtαEtγ,
balanced growth with constant income shares is possible even if E comes from a source in …xed supply (e.g., oil);
oil use should be (exponentially) decreasing over time;
and oil use is independent of technology, population growth, etc.!
The Cobb-Douglas framework still used by Nordhaus and Stern.
Previous research
A large literature on natural resource economics was triggered by the oil price shocks in the 1970s: Stiglitz, 1974, 1980; Solow, 1974;
Dasgupta and Heal, 1974.
Theoretical, not quantitative.
Abstracted from (endogenous) technological progress.
Showed that with Cobb-Douglas production function, Yt =AtL1 α γKtαEtγ,
balanced growth with constant income shares is possible even if E comes from a source in …xed supply (e.g., oil);
oil use should be (exponentially) decreasing over time;
and oil use is independent of technology, population growth, etc.!
The Cobb-Douglas framework still used by Nordhaus and Stern.
Previous research
A large literature on natural resource economics was triggered by the oil price shocks in the 1970s: Stiglitz, 1974, 1980; Solow, 1974;
Dasgupta and Heal, 1974.
Theoretical, not quantitative.
Abstracted from (endogenous) technological progress.
Showed that with Cobb-Douglas production function, Yt =AtL1 α γKtαEtγ,
balanced growth with constant income shares is possible even if E comes from a source in …xed supply (e.g., oil);
oil use should be (exponentially) decreasing over time;
and oil use is independent of technology, population growth, etc.! The Cobb-Douglas framework still used by Nordhaus and Stern.
Previous research
A large literature on natural resource economics was triggered by the oil price shocks in the 1970s: Stiglitz, 1974, 1980; Solow, 1974;
Dasgupta and Heal, 1974.
Theoretical, not quantitative.
Abstracted from (endogenous) technological progress.
Showed that with Cobb-Douglas production function, Yt =AtL1 α γKtαEtγ,
balanced growth with constant income shares is possible even if E comes from a source in …xed supply (e.g., oil);
oil use should be (exponentially) decreasing over time;
and oil use is independent of technology, population growth, etc.! The Cobb-Douglas framework still used by Nordhaus and Stern.
Previous research
A large literature on natural resource economics was triggered by the oil price shocks in the 1970s: Stiglitz, 1974, 1980; Solow, 1974;
Dasgupta and Heal, 1974.
Theoretical, not quantitative.
Abstracted from (endogenous) technological progress.
Showed that with Cobb-Douglas production function, Yt =AtL1 α γKtαEtγ,
balanced growth with constant income shares is possible even if E comes from a source in …xed supply (e.g., oil);
oil use should be (exponentially) decreasing over time;
and oil use is independent of technology, population growth, etc.! The Cobb-Douglas framework still used by Nordhaus and Stern.
Previous research
A large literature on natural resource economics was triggered by the oil price shocks in the 1970s: Stiglitz, 1974, 1980; Solow, 1974;
Dasgupta and Heal, 1974.
Theoretical, not quantitative.
Abstracted from (endogenous) technological progress.
Showed that with Cobb-Douglas production function, Yt =AtL1 α γKtαEtγ,
balanced growth with constant income shares is possible even if E comes from a source in …xed supply (e.g., oil);
oil use should be (exponentially) decreasing over time;
and oil use is independent of technology, population growth, etc.!
The Cobb-Douglas framework still used by Nordhaus and Stern.
Previous research
A large literature on natural resource economics was triggered by the oil price shocks in the 1970s: Stiglitz, 1974, 1980; Solow, 1974;
Dasgupta and Heal, 1974.
Theoretical, not quantitative.
Abstracted from (endogenous) technological progress.
Showed that with Cobb-Douglas production function, Yt =AtL1 α γKtαEtγ,
balanced growth with constant income shares is possible even if E comes from a source in …xed supply (e.g., oil);
oil use should be (exponentially) decreasing over time;
and oil use is independent of technology, population growth, etc.!
The Cobb-Douglas framework still used by Nordhaus and Stern.
What does that data say?
With the Cobb-Douglas production function, EP
Y should be constant.
Y - Real GDP, (BEA).
E - Fossil energy use (oil + coal + natural gas), (EIA).
P - Fossil fuel composite price (EIA, 2008, Table 3.1).
Is this true in the data?
A few key features of data: 1
Income share of fossil fuel highly variable in short and medium run and correlated with price.
19400 1950 1960 1970 1980 1990 2000 2010
0.02 0.04 0.06
The Factor Share of Energy in the U.S. Economy
19400 1950 1960 1970 1980 1990 2000 2010
1 2 3 4
Fossil Fuel Price (composite)
A few key features of data: 2
As noted above, energy use should fall at a rate given by the subjective discount factor.
19401 1950 1960 1970 1980 1990 2000 2010
1.5 2 2.5 3 3.5 4 4.5 5
Energy Consumption in the U.S.
Thus: the quantity data ‡atly contradicts CD too.
A few key features of data: 2
As noted above, energy use should fall at a rate given by the subjective discount factor.
19401 1950 1960 1970 1980 1990 2000 2010
1.5 2 2.5 3 3.5 4 4.5 5
Energy Consumption in the U.S.
Thus: the quantity data ‡atly contradicts CD too.
A few key features of data: 2
As noted above, energy use should fall at a rate given by the subjective discount factor.
19401 1950 1960 1970 1980 1990 2000 2010
1.5 2 2.5 3 3.5 4 4.5 5
Energy Consumption in the U.S.
Estimating the elasticity
Hassler, Krusell and Olovsson (2011) set up a more general production function and estimate the elasticity.
Finding: the elasticity between energy and capital is very close to zero.
Hence, no substitution in the short run.
Estimating the elasticity
Hassler, Krusell and Olovsson (2011) set up a more general production function and estimate the elasticity.
Finding: the elasticity between energy and capital is very close to zero.
Hence, no substitution in the short run.
Estimating the elasticity
Hassler, Krusell and Olovsson (2011) set up a more general production function and estimate the elasticity.
Finding: the elasticity between energy and capital is very close to zero.
Hence, no substitution in the short run.
Energy-saving technology
In the estimation process, we are able to derive a measure of energy-saving technological change.
1940 1950 1960 1970 1980 1990 2000 2010
0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6
•←The first oil shock AE
Regression lines
Figure:
We have a kink around the time of the …rst oil-price shock: an endogenous response?!
Energy-saving technology
In the estimation process, we are able to derive a measure of energy-saving technological change.
1940 1950 1960 1970 1980 1990 2000 2010
0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6
•←The first oil shock AE
Regression lines
Figure:
Smooth and increasing with a mean growth rate of 1.51%.
We have a kink around the time of the …rst oil-price shock: an
Co-movements of the technology trends
1940 1950 1960 1970 1980 1990 2000 2010
0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6
A AE
Figure:
The productivity slowdown coincides with a faster growth in the energy-saving technology - the two series are mirror images.
Looks like a case for directed technical change – scarce R&D resources creates a negative trade-o¤.
Co-movements of the technology trends
1940 1950 1960 1970 1980 1990 2000 2010
0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6
A AE
Figure:
The productivity slowdown coincides with a faster growth in the energy-saving technology - the two series are mirror images.
Looks like a case for directed technical change – scarce R&D
A model with a low elasticity can match the data
Hassler, Krusell and Olovsson (2011) set up a model with a low elasticity and where K , E , A and AE are all endogenous.
We show that the model can match the data on shares and oil use (peak oil).
Implication for policy?
So far, we have not analyzed optimal policy in the model, but the
…ndings suggest that
high prices/taxes will have limited e¤ects on oil use in the short run.
In the long run, more e¢ cient technologies are developed and energy use goes down.
The cost is that other technologies that improve on capital/labor will grow at a slower pace.
A related study
Aghion, Acemoglu, Burstyn och Hemous (2011) also consider a model with clean and dirty energy.
They …nd that the technological process is crucial.
Immediate action is needed to avoid a disaster.
The mechanism comes from path dependence in the technological process.
Conclusions
Policy recommendations must be based on models that are consistent with the real world.
This is not always the case.
The elasticity between energy and capital is for instance lower than what is generally assumed.
