Limiting Green House Gas emissions:
an economist’s perspective
Thomas-Olivier Léautier ([email protected])
with Claude Crampes ([email protected])
Les Houches, February 2014
Outline
1. Clean Energy Policy for Europe
2. Basic microeconomics for externalities
3. The European Emission Trading System
4. Microeconomics for cap-and-trade
2
EU Green-House Gas emissions towards an
80% domestic reduction (100% =1990)
100%
80%
100%
Power Sector
80%
Current policy
60%
Residential & Tertiary
60%
Industry
40%
40%
Transport
20%
20%
Non CO2 Agriculture
Non CO2 Other Sectors
0%
1990
2000
2010
2020
2030
2040
0%
2050
Source: European Commission, “A Roadmap for moving to a competitive low carbon economy
in 2050”, March 2011
3
2. Basic microeconomics for externalities
price
Supply
p market
Equilibrium
Demand
quantity
0
q market
4
GHG emissions as a negative externality
Negative externality associated with GHG emissions:
●
emitters do not face the full social costs of emissions, including their
impact on the environment (global warming).
Without intervention, the market would emit excessive pollutants
Source: IPCC (2007)
5
Negative externality and market failure
price
cost of the negative
externality
social marginal cost
supply
=
private marginal cost
optimum
equilibrium
demand
0
qoptimum qmarket
quantity
6
A series of complex issues
●
Physics (climate science):
●
●
Engineering
●
●
What is the impact of temperature increase?
What technical progress can be expected?
Economics:
What is the cost of temperature increase?
● What is the cost of decarbonization?
●
What weight for future generations versus current ones?
● How to split the burden between developed and developing
countries? between industries?
●
●
How to limit opportunistic behavior?
7
Controlling GHG emissions: what is the right
method?
• Overall objective: minimizing the cost of reducing carbon
emissions
o Set of policies that directly address the market failures associated with
climate change, and only intervene where market failures are present
o Technology- and sector-neutral approach to carbon abatement:
carbon reduction in sectors which have the lowest cost of reducing
emissions
• Potential policies
o Direct regulation: Command & Control, prohibition, quotas,
standards…
o Incentive regulation: carbon pricing (cap-and-trade, carbon taxes),
subsidies and R&D incentives
8
How to create a carbon price?
i) tax
price
social marginal cost
modified
private marginal cost
supply
=
private marginal cost
optimum
equilibrium
carbon
tax
demand
0
qoptimum qmarket
quantity
9
How to create a carbon price?
ii) tradable permits
social marginal cost
price
volume
cap
supply
=
private marginal cost
Optimum
price of the
permit
price of the good
demand
0
qconstrained
quantity
10
Market vs. tax
• Principle of responsibility: the polluter must pay (article 174-2
of the Treaty); Is the producer or the consumer the true
polluter? Is cost pass-through acceptable?
• Carbon tax : Who is in charge? How is it calculated? Who
receives the cash? What to do with revenues?
• Tradable permits : Who decides? How many allowances? If
given for free, to whom? If sold, who benefits from sale?
• Theory (Weitzman, 1974): quantity control is more efficient
than price control when supply is more inelastic than demand
11
Price vs. quantity regulation
price
social marginal cost
Net
Surplus
p*
Average demand
0
q*
quantity
12
Welfare loss under quantity regulation
price
social marginal cost
Surplus loss under quantity
regulation
p*
Realized demand
Average demand
0
q*
q**
quantity
13
Welfare loss under price regulation
price
social marginal cost
Welfare loss under price
regulation
p*
Realized demand
Average demand
0
q*
q**
quantity
14
Price vs. quantity regulation 2
•Political economy: potential for regulatory capture
produces first-order effects
•Concerning
CO2
emissions,
Directive
2003/87/CE has set the framework: the EU-ETS, a
cap-and-trade system
15
3. The European Emission Trading System
• Directive EU ETS (European Emission Trading Scheme) in 2003
before the commitment from the Kyoto protocol.
• Three compliance phases
• Now 28+3 heterogeneous States participate
1 Jan. :
ETS Phase I
2005
1 Jan. :
ETS Phase II
2007
Feb. : Kyoto
protocol comes
into force
2008
1 Jan. : beginning
of first Kyoto
protocol period
3x20
1 Jan. :
European Objectives
ETS Phase III
2012
2013
2020
Dec. : end of first
Kyoto protocol
period.
16
Cap-and-trade principles
●
●
Binding cap is set on emissions during a given period
Emission permits are allocated to polluters:
o
auction or free allocation based on grandfathering or benchmarking
www.eex.com/en/Market%20Data/Trading%20Data/Emission%20Rights/EU%20Emission%20Allowances%20%7C%20Spot
●
Emission permits can be traded (wholesale or, mainly, OTC):
o
o
o
●
Regardless of the initial allocation (if no transaction costs), trading allows for an
optimal distribution of abatement efforts across sectors and countries (Coase
principle)
The initial allocation of permits only has a wealth effect
If the allocation is auctioned, second hand market is just for efficient adjustment
Polluters not allowed to emit more than initial allocation +
permits bought on the market; otherwise, they pay a penalty.
17
European Environmental Policy: 2013-2020
The ETS Directive (2009/29/EC):
●
●
●
From 2013 onwards (Phase III), emission allowances in the
ETS will be reduced by 21% below their 2005 levels by 2020
Full auctioning for the power sector, and a gradual phasing out
of free allowances for other sectors
The ETS is also set to be expanded from 2013, to also include
the aviation sector. But …
18
Allocations by sector
19
19
Allocations by country
20
Total allowances
21
Flexibility
Banking
Emissions permits can be used in periods subsequent to the one in
which they were allocated. Inter or intra-phase?
– in Phase I, only intra-phase
– now also interphase (Phase II => Phase III)
Borrowing
Allows regulated emitters to use part of their future allocations to cover
their present emissions
– de facto allowed intra-phase (February 28 => April 30)
Credits offset:
- Clean Development Mechanism
- Joint Implementation
22
Low carbon price
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
Prix spot 2005-2007
Prix spot (depuis 2008)
23
4. Microeconomics for cap-and-trade
 max
x, y
pQ( x)  wx  g ( y )
x : input = gross emissions, unit price w
Q(.) : output
Q '  0, Q ''  0, unit price p
y : abatement effort
g (.) : abatement cost ,
g '  0 , g"  0
e  x  y : residual polluting emissions
def
w
 FOC: pQ '( x)  w  demand for input X ( w, p)  Q ( ) and y  0
p
' 1
 Assume the emissions without constraint eo  X ( w, p)  e ,
where e is the social optimum level.
24
24
Permits
• The authority limits to e the emissions allowed (cap)
and open a permits exchange (trade)
• qa  0 demand (qa  0 supply) of allowances
pa unit price of allowances
pQ( x)  wx  g ( y)  qa pa s.c. x  y  e  qa (  )
• max
x, y ,q
a
• FOC
x:
pQ '( x)  w    0
y :  g '( y )    0 ( 0 if y  0)
qa :  pa    0
25
Trading
0
 When pa  0, we have   0. Then qa  x  y  e 
is the demand for rights derived from the firm’s optimal
production.
 Assume first that g '(0)  pa ; then y  0. From the two other FOC,
we obtain the gross demand for emitting CO2 def
w  pa
pQ '( x)  w  pa  0  X ( w, pa , p)  Q '1 (
).
p
 The firm is a net demander of allowances if pa is small,
a net supplier otherwise.
0
qa ( w, pa , p )  X ( w, pa , p )  e 
26
abatement effort and market of
allowances
 Assume now g '(0)  pa ; the firm fixes y so that g '( y)  pa .
def
Let Y ( pa )  g '1 ( pa ); it is increasing in pa since g ''  0.
def
• We then have q (w, p , p) = X (w, p , p) -Y( p ) - e
a
a
a
a

where Y ( pa )  0 for pa  0, pa

pa
qa (w, pa , p)
qa (w, pa , p)
X ( w, pa , p)
pa1
effort quota
traded
permits
e
e, x
qa ( w, 0, p)
27
Equilibrium
 The market equilibrium is reached at pa* such that
 X (w, p , p) Y ( p )   e
or å q (w, p , p) = 0
i
a
i
i
i
ai
a
i
i
a
i
For two « price-takers »
qa1 (w, pa , p) pa
pa*
qa2 (w, pa , p)
pa2
pa2 < p < pa1
*
a
pa1
åq
ai
At equilibrium, 2 is a seller
and 1 is a buyer.
i
qa1 , qa 2
28
comparative statics
The market equilibrium varies with
- product(s) price(s) p: electricity, aluminium, steel, etc.
- input(s) price(s) w: coal, natural gas, fuel, etc.
- the initial total endowment e ,
not the allocation (e1 ,..., en ) : R. Coase
- the shape of the functions of abatment cost g (.)
29
Paying for allowances
*Assume that firms have to pay s for each ton initially allowed
max
x,qa
()
pQ(x) - wx- paqa - g y - se s.c. x- y £ e + qa
* The marginal conditions remain the same as s et e are exogeneous.
* Risk of foreclosure if the global "tax" s e is too high.
30
Auctionning allowances
• Under the first Directive, only four countries have used the
possibility to sell (at most 5 %) allowances : Denmark (5 %),
Hungary (2.5 %), Lithuania (1.5 %) and Ireland (0.75 %).
• Under the 2009 Directive, it is 100% mandatory for the
electricity producers from 2013 on. Partial obligation in the
other industries.
• Then, to produce output q, a firm can now obtain permits from
free allowances, e
auctionned permits, a
(cost 0)
(unit cost s)
abatement effort, y (total cost g ( y ))
traded permits, qa
(unit cost pa )
31
Conclusions
• For the EU authorities, it takes (at least) three tools to reach
objective:
― one for cleaning (Directive 2009/29/EC: mandatory ETS),
― one for greening (Directive 2009/28/EC: optional green
certificates, or FIT, or green potfolio, or …),
― one for saving (Directive 2012/27/EU: optional white certificates, or
energy efficiency, or load-shedding, or demand response, or …).
• Actually:
― as the objective is to cut GHG emissions, one tool is sufficient
― combining several tools produces negative side-effects.
32
An economist perspective

Cap and trade for CO2 is a right answer because
o
o
o
o

it fixes a negative externality;
it sends a scarcity signal;
it allows firms to adjust volumes;
it (now) generates public revenues.
Independent quantitative targets for energy saving and
renewables are wrong answers because
o they are viewed as genuine objectives instead of mere means;
o they increase the cost of reaching the CO2 target;
o they require large amounts of red tape and (distortive) State
aids.
33
Appendix
EU-ETS timeline
end of year N
beginning of year N
double allocation period
1st Jan.
28 Feb.
Year N allocation on
installations accounts
in their national
registry.
30 March
Publication of year N-1
emissions by the EC
30 April
Installations submit
their verified
emissions for year
N-1 to the national
authority.
15 May
31 déc.
Installations surrender the
allowances covering their N-1
emissions in the national
authority.
35
w
w
pQ '( x)  pa
pQ '( x)  pa
•
w
pQ '( x )
pQ '( x )
w
• •
•
••
pa
x, e
o
X ( w, pa , p) e e
net supply
pˆ a
X ( w, pa , p)
x, e
net demand
net supply
qa ( w, pa , p)
eo
qa ( w, pa , p)  0
qa ( w, pa , p)  0
pa
e
pa
X ( w, pa , p)
qa ( w, pa , p) X ( w, pa , p)

pa
pa
net demand
X ( w, pˆ a , p)  e
pˆ a : switching threshold
qa (w,0, p) e X ( w, 0, p ) qa , x
36
Timeline
Firms will be active on both the initial sale and on the permits
exchange only if there is some randomness (on p and/or w)
auction a
quota
checking
e
time
p and w certain
max
a

trading (choice of 𝑞𝑎 )
E p , wu  sa  max  pQ( x)  wx  pa qa
x , qa
s.c.
x  e  a  qa 

where u '  0, u ''  0.
37
* Ex post, knowing p and w we have that
qai  xi  ei  ai

 
'
and pQi ( xi )  w  pa  0
pQi' (qai  ei  ai )  w  pa
 w  pa
* Individual net demands qai  Q 
 p
' 1
i

  ei  ai ,

are agregated for all i to give the price on the permit exchange




pa    ei  ai , w, p  and the net demands qai    ei  ai , w, p  .
 i

 i

* Remark : If most firms are net suppliers ( xi  ei  ai ) [for example
because w is larger than expected] there is no pa  0
such that  qai  0 : the equilibrium price is nil.
i
38
* We still have to determine how much to buy in the initial auction
max
ai
E p , wui   sai  pQi  qai (.)  ei  ai   w  qai (.)  ei  ai   pa (.)qai (.) 
or
max E p , wui  pQi  qai (.)  ei  ai    s  w ai   pa (.)  w qai (.)   wei
ai
* The FOC is
(with
pa
 0)
ai

 qai
'
E u (.)  pQi (.) 1 
ai


'
w, p i

qai
  ( s  w)  ( pa (.)  w)
ai


  0

39
qai
 1, we get Ew, p ui' (.)  pa (.)  s   0
ai
to determine the initial demand for permits ai ( s).
* Since
* Equilibrium is given by
 a (s)  supply.
i
i
• Remark 1: s 
Then
•
•
Ew, pui' (.) pa (.)
Ew, p ui' (.)
if risk neutral, i buys on the initial auction only if s  E p (.)
w, p a
if risk averse, we have s 
Ew, pui' (.) pa (.)
'
w, p i
E u (.)
 Ew, p pa (.)
meaning that i is ready to pay a risk premium.
40
* Remark 2:
• If the initial auction of permits is not followed by trading possibilities, it
is a private value auction : each firm bids a price only based on its own
characteristics.
• Opening an ex post exchange for permits transforms the auction into a
'
common value auction : Qi depends on qai that depends on the total
number of allowances and the technical characteristics of all obliged
firms.
41
Dynamic opportunism
• During the first round (2005-2007), the EC has announced
that future quotas would not be based on the observed
performances of the current round, to reduce opportunism.
• Actually, “it is useful to learn from the most recent data”,
• Finally, the expected individual emissions for 2008-2012
have been based on declared emissions of 2005 multiplied
by an expected growth rate until 2010.
• What is the risk?
42
Grandfathering: internalizing the
review rules
T
max
( x , qa )


  p Q ( x )  w x  pa qa 
 0
gives FOC
s.t. x  e  qa

0   ( pQ ' ( x )  w  pa )  
 1
  0,..., T .
d e 1
  2 d e  2 d x 1 d e 1

 ...
dx
dx 1 d e 1 dx
With just a one-period effect
d e 1
pQ ( x )  w  pa  
 w  pa
dx
'
d e 1
As
 0 and Q'' ( x )  0, the firm chooses a larger x
dx
which means larger emissions than if e 1 is fully exogeneous.
Not the case when allowances are auctioned.
43