1
October 3, 1999
Large-Country Effects in International Emissions Trading:
A Laboratory Test*
by
Björn Carlén
Department of Economics, Stockholm University, SE-106 91 STOCKHOLM
Abstract
The experiment mimics carbon emissions trade among twelve industrialized
countries during the end of a five-year-long trading period when traders are likely
to have nearly full information about the underlying net demand. Trade is
assumed to be governed by so-called double-auction rules. The hypotheses are (i)
that the presence of large countries would not prevent trade from being efficient
and (ii) that large countries would not be able to influence price levels to their
advantage. The findings support the first hypothesis but are inconclusive
regarding the second, although they illustrate that a large country may not be able
to sustain favorable prices.
Key words: Carbon emission trading; Market power; Experimental economics
JEL Classification: Q28; L1; C9
* Financial support from the Nordic Council of Ministers and the Jan Wallander
and Tom Hedelius Foundation is appreciated. The author is grateful to Anders
Andersson, Peter Bohm and Carsten Helm for valuable comments on earlier
versions of this paper, Astri Muren and Lars Vathrik for assistance in running the
tests, and Lars Kristensson for computer assistance.
2
1.
Introduction
International emission trading in the form of a tradable emission quota (TQ) market
appears to be the most promising policy measure within the field of international
climate change policy. It is well known that a well functioning TQ market minimizes
the costs of achieving a given aggregate emission target if transaction costs are low,
and that such a market is capable of doing so for a wide range of initial quota
allocations1, Montgomery (1972). Since emissions of the single most important
greenhouse gas, carbon dioxide (or carbon) from combustion of fossil fuels, can be
monitored quite easily by keeping track of countries’ use of fossil fuels, IPCC (1996)
and Bohm (1998), the advantages of international emission trading seem possible to
realize at least for the case of carbon emissions. However, several countries, in
particular developing countries, have been skeptical towards such a policy, a
standpoint possibly caused by concerns regarding market power. Fear that market
power would pose a problem seems to be based on expectations that (some) countries
would engage in intergovernmental emission quota trade rather than letting their firms
trade internationally. If so, large countries might dominate the market outcome.2
So far, the policy discussion about market power in emission trading of this kind
appears to have been influenced mainly by standard economic theory, see e.g., Hahn
(1984), Tietenberg (1985) and Westskog (1996). This kind of analysis predicts that
dominant sellers (buyers), in order to maximize their profits, withhold supply (demand)
from the market, a behavior that inevitably results in lower efficiency and smaller trade
gains for buyers (sellers) than had the market been perfectly competitive. However,
these results are derived under the assumption that emission trading is governed by a
set of trading rules (a “trading institution”) that requires all trade to be conducted
simultaneously at a uniform price and restricts agents to choose only what quantity to
supply or demand. It is not predetermined that trade with carbon emission quota units
would be governed by such an institution. On the contrary, the discussion so far
1
The term emission quotas as it is used here corresponds to the term “assigned amounts of national
the Kyoto protocol.
2
In the case of a climate treaty only comprising industrialized countries as would be the case if the
Kyoto protocol were to enter into force, Russia and Ukraine are likely to act as major sellers while the
US, which stands for about one half of the aggregate OECD carbon emissions, is likely to turn out as
a large buyer. A global treaty would add large countries such as China and India to the list of expected
sellers.
3
indicates some version of the continuous double auction (DA), or as it is often called,
an open-outcry market, as the prime candidate, Sandor et al (1994). Under DA rules
trade is sequential and traders can state bids and asks or accept (part of) other traders’
bids and asks. Hence, different transactions can be concluded at different prices.
Versions of the DA institution have been chosen to govern trade on exchanges such as
the New York stock exchange and the major Chicago commodity exchanges, Friedman
(1993).
Despite the importance and long history of DA-based markets no general theory
regarding behavior under DA rules has yet emerged.3 However, the experimental
literature suggests that standard economic theory regarding market power may not be
relevant for the case of carbon emission quota trade governed by DA rules. A large
number of stylized laboratory experiments indicate that the DA institution tends to
produce efficient outcomes also in cases where the market consists only of few traders
and to make it difficult even for a monopolist to attain the monopoly profit level, see
Plott (1989), Davis and Holt (1993) and Holt (1995). This is not to say that certain
effects of market power cannot arise in DA markets. It has been observed in stylised
DA experiments where trade (with a quite small number of units) has been based on
private information that relative size matters, e.g., Smith (1981). In the context of
tradable emission permits (the domestic version of emission trading) Ledyard and
Szakaly-Moore (1994), Brown-Kruse et al (1995), Godby (1998) and Muller et al
(1999) have reported stylised laboratory experiments in which monopolists
(monopsonists) were able to influence price levels to their advantage, which in some
instances seems to have reduced the level of market efficiency. 4, 5
So, although the DA institution appears to be better than other institutions in
preventing market power from emerging, the mixed results of laboratory experiment
3
Some theoretical contributions have been made by Wilson (1987), Friedman (1991), Easley and
Ledyard (1993) and Gjerstad and Dickhaut (1995).
4
Misiolek and Elder (1989) have shown theoretically that it may pay a dominant firm may to act
strategically on the emission permit market (by hoarding permits) in order to influence conditions on
a downstream market and thereby increase its overall profits. Laboratory experiments have shown that
such behavior can reduce the performance of emission trading substantially even when it is governed
by DA rules, e.g., Godby (1997). However, this kind of strategic behavior is not likely to arise in the
context of international emission trading.
5
For a review of experimental literature on market power in domestic emission trading see Godby et
al (1998).
4
regarding domestic emission trading indicate that dominant traders may be able to
influence price levels to their advantage also under DA rules.
Experimental work regarding international emission trading is so far quite limited.
One example is Bohm (1997) who conducted an emission reduction trade experiment
with real-world relevant decision makers and marginal abatement cost functions for the
case of bilateral emission trading among four Nordic countries. In this experiment
nearly all potential trade gains were realized, a result corroborated by a set of pilot
(laboratory) tests preceding it, Bohm and Carlén (1999). Moreover, Hizen and Saijo
(1998) have conducted a set of laboratory tests of emission trading among Russia, the
Ukraine, the US, Poland, EU and Japan, comparing the performance of bilateral trade
and trade governed by DA rules under different information treatments. They report
that both institutions yield high efficiency while prices tend to converge to the
competitive price level only under DA rules.
The purpose of the experiment presented here is to test the effects large countries
may have on the outcome of international carbon emissions trading when trade is
governed by DA rules. In principle, several options are open for the design of such an
experiment. However, getting countries to volunteer in performing real emission
trading on an experimental basis is hardly an alternative. Since the situation to be
tested here involves a quite large number of countries (in order to motivate the use of
an organized DA market), also the approach of conducting hypothetical emission
trading (but with relevant trader incentives) under as field-like conditions as possible as
in Bohm (1997) seems unsuitable. The remaining approach and the one taken here is to
conduct a laboratory experiment using demand and supply conditions as well as an
information structure that are supported by empirical studies.6
The testbed used here is based on the quota allocation proposed in the Kyoto
protocol, UN (1997), and on estimates (when available) of marginal abatement cost
(MAC) functions of some industrialized countries, IPCC (1996), EU (1996) and
6
Although far from real carbon emission quota trading, this kind of laboratory experiments can
provide relevant information about behavior on a TQ market for carbon emissions. An indication of
this is given by the above mentioned study by Bohm (1997) and the preceding set of laboratory tests
that used similar demand and supply conditions in order to identify a well-functioning experimental
design, Bohm and Carlén (1997). A comparison between these pilot tests and the full-scale
experiment revealed no large differences between emission trading by students facing monetary
incentives in a laboratory and hypothetical emission trading by real-world relevant subjects that were
given strong non-monetary, but trade relevant, incentives.
5
Capros et al (1997). The experiment intends to mimic an assumed situation in which a
group of twelve industrialized countries trade carbon emission quota units on a DAbased market where demand and supply conditions are such that a single country (the
U.S.) has a dominant role in that it would buy 90 percent of the competitive supply.
More specifically, it is designed to mimic trade during the end of a five-year long
carbon emission budget/trading period, a time at which uncertainty and information
asymmetries would be likely to be small or negligible. 7 This is the crucial part of the
trading period when compliance with the Kyoto protocol is to be ascertained if the
Protocol would come into force. Moreover, trade during the end of the trading period
determines whether or not trade during the whole trading period will yield an efficient
allocation of carbon emission abatements.
The paper is organized as follows. The next section (Section 2) outlines the trading
situation that countries engaging in carbon emission quota trade would be likely to face
and discusses some implications for large countries’ ability to exert market power.
Section 3 describes the experimental design. Section 4 presents the results of the
experiment. Conclusions are offered in a final section.
2.
The Trading Situation
The point of departure for the analysis presented below is a climate treaty that allots to
each country an emission quota for the sum of its carbon emissions over a five-year
long emission budget period (e.g. 2008-12 as in the Kyoto protocol), and allows the
signatories, i.e., the governments, to trade quota units with each other.8 Trade is taken
to be governed by DA rules. A TQ market covering countries from different regions of
the globe, e.g., US, Europe and Japan, might be taken to be open all hours, Sandor et
7
A reason for focusing on this situation and not on the whole five-year long trading period is the
difficulties of inducing uncertainty and controlling for differences in subjects’ risk attitudes while
constructing an experimental situation that subjects are likely to grasp.
8
Due to uncertainty, the kind of climate treaty discussed here might allow countries to bank and
possible also borrow (up to some limit) emission quota units to/from the next five-year budget period.
Moreover, countries would also be likely to have the opportunity to trade on futures markets, i.e.,
trade quota units valid for the next budget period. These aspects of emission quota trade are not dealt
with here.
6
al (1994). Hence, after the opening trade at the beginning of the budget period, which
might be conducted under specific rules (i.e. as a call market in the same fashion as at
the N.Y. stock exchange), there would be no recurrent opening trade.
A TQ treaty allows a country to emit more than its assigned amounts of emissions
(its initial quota level) as long as it buys emission quota units from other countries.
Through this kind of transaction a seller country transfers part of its emission budget
to buyer countries, whereby the aggregate emission level of all the countries remains
the same. Figure 1 illustrates the underlying net demand of a country with a strictly
binding emission quota (e^ ) and a MAC function that increases and is convex in
emission abatements. The country is committed to be accountable for an emission
reduction equal to the difference between its business-as-usual level (e0) – BAU for
short – and e^ .9 If the country would accomplish this emission reduction solely by
domestic abatements it would incur abatement costs equal to the area e^e0MAC(e^).
/Figure 1 about here/
The country can reduce its cost of compliance (make trade gains) by either replacing
domestic “high cost” abatements with “low cost” abatements via purchase of quota
units from other countries or selling quota units at prices higher than its MACs. That
is, for prices below (above) MAC(e^) the country has a positive (negative) underlying
net demand for quota units. Hence, countries with strictly binding emission quotas will
be net sellers for sufficiently high price levels and net buyers for sufficiently low prices.
(The concept of net demand is defined in relation to the country’s “target level”, i.e.,
its initial quota (e^) plus conducted net purchases of emission quota units. Figure 1
illustrates the underlying net demand of a country at the beginning of the trading
period when the target level equals e^.)
Given the high values at stake in carbon emission quota trade, countries would have
strong incentives to gather information about expected MAC functions of other
countries prior to emission quota trading. It is unlikely that individual countries would
be successful in concealing from other countries large parts of its abatement
9
The fact that individual countries would be likely to conduct some abatements also in the absence of
a climate treaty, and hence, that a country’s BAU emission level is not equal to the point where MAC
curve crosses the x-axis, is suppressed here.
7
opportunities (and the costs of these). Therefore, it is assumed here that participating
countries would have roughly common information about each others’ (and the
aggregate) expected underlying net demand relationships for the budget period as a
whole. This implies that countries would have more or less common expectations
about the efficient price level (P*). As new information becomes available, countries
would be able to update their expectations, whereby towards the end of the five year
long budget period they are likely to have even better information about everybody’s
MAC functions. Hence, when, say, 4-6 months of the trading period remains, the
countries would be able to estimate P* with good approximation.
A country that sees price levels materialize that deviate substantially from its current
expectations would be likely to update its expectations regarding future price levels or,
as long as its trading budget so allows, engage in speculative trade, or some
combination thereof. The option to speculate implies that countries expected to act as
large net sellers may for some time appear as buyers on the market and vice versa.
The trading situation outlined above has several implications for the ability of
dominant countries to exert market power, three of which will be highlighted here.
(1) On markets of emission trading, in contrast to many other commodity markets,
countries may shift trader roles when prices change. Thus, even if a buyer country
dominates the market for prices around P, there exist prices asked or bid significantly
below P which the dominant country would like to see materialize, other things equal,
but which would turn “former” sellers into buyers, hence creating competition for the
original buyer (mutatis mutandis for an originial seller). A TQ market for carbon
emissions would, therefore, not allow a country to remain in an equally dominant
position when prices become increasingly favorable to him and unfavorable to others.
(2) From the assumption that trade will be governed by the DA institution it follows
that final trades can be expected to occur at or near P*. Since trade under DA rules is
sequential, a dominant agent can exert market power either by price discrimination or
by withholding demand/supply from the market in order to influence prices, or some
combination thereof. If a monopsony (monopoly) would succeed in conducting price
discrimination of the first degree, transactions would be concluded at over-time
ascending (descending) prices. Since revenues of earlier transactions are unaffected by
additional trade, there exist incentives for the monopsony (monopoly) to trade until the
price level approaches P*. Similarly, if a monopsony (monopoly) according to standard
8
theory of market power would hold back demand (supply) to competitive sellers
(buyers), in an attempt to reduce (increase) prices, there would come a time when a
temporarily reduced transaction volume reveals that both buyer(s) and seller(s) have
reservation prices for additional units that make further trade profitable for both
parties. The prices for such trade will eventually approach P*. So, even if a dominant
agent in a DA based market is able to influence prices, it is unlikely to cause prices of
all transactions, and the net trade volume, to deviate from the efficient level, see Plott
(1989) and Holt (1995) for a discussion about experimental results. Instead, the likely
outcome of a successful monopsonist (monopolist) would be prices that start out from
below (above) P* and converge to the efficient level.
(3) However, since experienced traders would be aware of the outcome just
mentioned and since these traders would be equipped with more or less common
information, thereby having more or less common expectations, early prices are not
likely to deviate significantly from P*.10
An empirical regularity in DA experiments with private information regarding
demand and supply conditions, a so-called convergence bias, deserves attention here.
By convergence bias is meant the tendency of prices to converge towards P* from
below/above if aggregate producer surplus (evaluated at the competitive outcome)
exceeds/falls short of aggregated consumer surplus, Smith and Williams (1982).11
Moreover, “excesses of producer over consumer surplus are found to have a more
pronounced effect on the sequence of contract prices”, i.e., buyers tend to make better
deals than sellers. Smith and Williams regard this as something that may be due to the
fact that experimental subjects have more experience as buyers than as sellers. Since
these phenomena have been observed in experiments where the total number of units
have been distributed evenly among the subjects, their effects on prices cannot be
attributed market power. In the experiment presented below, the producer surplus in
the fully competitive case exceeds the consumer surplus by 20-96%. Hence, if trade
would be conducted under private information only, it is likely that we would observe
10
On a real-world continuous market, new prices are likely to be affected also by past prices. Thus,
approaches towards a new equilibrium price are likely to differ if the past equilibrium price was
higher or lower. This aspect is not explicitly taken into account here.
11
A related observation is that prices of initial contracts (usually concluded by the seller with the
lowest reservation price and the buyer with the highest reservation price) tend to be such that trade
partners divide the available surplus equally, Davis and Holt (1993). That is, initial prices tend to lie
close to the arithmetic mean of the maximum willingness to pay and minimum willingness to accept.
9
price paths converging to P* from below. That is, an outcome similar to the case
where the dominant buyer country had been successful in exerting market power.
However, as stated in the preceding paragraph, the convergence bias is not likely to
occur here. Instead, given the assumed information structure, the price path would be
largely independent of the relation of buyers’ and sellers’ surpluses at price P*.
To sum up, while standard economic theory predicts that the presence of dominant
traders on a TQ market will cause inefficiencies and change the distribution of trade
gains (in relation to the perfectly competitive outcome) to their advantage, the relevant
trading situation at the end of the budget period leads to the hypotheses (i) that the
presence of dominating countries will not prevent the market from being efficient, i.e.,
reaching the cost-effective allocation of abatements, and (ii) that dominant countries
will not be able to exert market power in the sense of establishing early prices that
deviate substantially from P*. Hence, it is hypothesized that the outcome of this kind
of end period trade will be close to the perfectly competitive outcome.
3.
Experimental Design
In order to mimic a situation of emissions trading among experienced traders, subjects
were recruited among higher undergraduate students in Economics or Finance at
Stockholm University and Stockholm School of Economics, with the following
exception. Since it was crucial for the testing of the influence of market power that
subjects playing the role of the dominant country would be confident in this role, these
subjects were selected amongst Ph.D. students in Economics that had proved their
negotiation skills in earlier laboratory experiments, here, those reported in Bohm and
Carlén (1997).
An often used approach of producing a certain information/expectation structure
has been to let subjects trade repeatedly. However, this approach is inappropriate here
since each trading period in the experiment has to be of substantial time length in order
to give subjects ample time to compute and exchange the desired, possibly large,
number of units. Given a time budget for each subject’s participation of a maximum of
four hours, the approach chosen was to give the subjects extensive written
information, prior to the experiment, about the context and the information structure
10
of the relevant trading situation. (The instructions are presented in Appendix A). A
summary of the experimental design is given below.
Information about the identities of the pioneering countries that are likely to engage
in real emission quota trading is not (yet) available. However, since much of the debate
of the market power issue relates to the US, it seems appropriate to include that
country. The approach taken here is to assume that twelve industrialized countries
would engage in carbon emission quota trade, the US, Japan and ten representing EU
(Belgium, France, Germany, Greece, Italy, the Netherlands, Portugal, Spain, Sweden
and UK.).12 Furthermore, since we can only guess what market structure would
emerge at the end of a five-year long trading period, it is assumed that when 4-6
months remain, these countries would be 20% from full compliance of a Kyotoprotocol-like environment. Combining this assumption with publicized estimates of
these countries’ expected BAU emission levels and assumptions about their MACs
yields a market structure with nine expected net sellers and three expected net buyers
where the main buyer (the U.S.) alone stands for approximately 90% of the efficient
net trade of approximately 400 (units) Mton CO2.13
The market institution used was the computerized multi-unit double-auction market
(MUDA) developed at Caltech. This institution was chosen since it allows subjects to
trade without being assigned predetermined buyer or seller roles.
Two days before the experiment, subjects received instructions including an
n introduction to the climate change policy issue,
n information about all countries’ expected underlying demand (MAC) functions
and the competitive outcome calculated on basis of this information,
n the trading rules,
n the show-up fee and
n a statement that they could earn considerable amounts of money by trading
12
The EU countries omitted are Austria, Denmark, Finland, Ireland and Luxembourg, all countries
which according to the Kyoto protocol and available estimates of BAU emission levels can be
expected to play a negligible role on this kind of market.
13
The Kyoto protocol comprises a requirement of emission trading being supplemental to domestic
actions, Article 17. However, the concept of supplementarity remains to be defined. Therefore the
experiment presented here abstracts from the requirement of trading being supplemental.
11
emission quota units on the behalf of the (unidentified) country they would
represent.14
The subjects were informed about the fact that the countries’ expectations about
their MAC functions could be updated and that they would receive, as private
information at the beginning of the experiment, the relevant MAC function for the
country they would be asked to represent. It was stressed that only small information
asymmetries were to be expected. Furthermore, in an attempt to mimic the situation of
experienced and well informed traders outlined in Section 2, the subjects were given
information about the expected outcome on a DA market, i.e., that marginal prices
were likely to end up near the efficient price level and that traders might engage in
speculative trade if current price levels would deviate substantially from the price level
they expected to prevail at the end of the trading period.
All subjects participated in a training program a couple of hours prior to the
experiment. This training program consisted of (i) an introduction to the computerized
double auction mechanism and the educational software included in MUDA, (ii) a
repetition of essential parts of the written instructions, and (iii) two training rounds in
order to make subjects familiar with trading on a MUDA-based market and with
interpreting the kind of underlying net demand functions that would be used in the
experiment.15
Once a subject had been assigned the role of a particular (unidentified) country,
he/she received as private information an updated version of that country’s MAC
function (the country’s underlying net demand function) and the size of that country’s
trading credit. No information was given whether other countries’ MAC functions had
been updated. (The distance between the expected and updated MAC curve of a
country was never larger than 4% of the expected curve, a fact that was not disclosed
to the subjects.) Since it was stated in the information common to all that countries
were of different sizes the subjects were likely to expect that they would have trading
14
In order to prevent the trading behavior from being influenced by knowledge about which countries
subjects represented, the countries were labeled country #0, #1, .... #11 during the experiment.
15
The first training round utilized monetary incentives. In this round subjects were assigned the role
of either a buyer or a seller with positive marginal valuations or costs for only a few units, information
that was private to each subject. In the second training round no monetary incentives were given and
the subjects were assigned the roles of traders. Here, trade was based on common information about
underlying net demand functions of the same type that would be used in the experiment. However,
here no trader dominated the market.
12
budgets of different sizes. As a matter of fact, each country’s trading budget was set so
as to correspond to the higher of the following two values, (1) its expected quota
import expenditures in the competitive outcome plus 30% of that amount and (2) two
percent of its GDP. Thus, all countries, including expected net sellers, were given a
trading budget that allowed them to engage in speculative trade from start.
Incentives: In order to mimic a situation in which traders representing countries on
a real TQ market for carbon emissions have incentives to try to maximize their
respective countries’ trade gains, subjects were paid a fraction of the trade gains they
attained for their countries. Since the underlying demand and supply conditions implied
a quite unequal trade gain distribution for the case where traders behaved as under
perfect competition (see Table 1 and Table 2 below), differentiated personal payoff
factors was used so as to produce a more even distribution of incentive payments. The
subject representing country #0 received SEK 6 per SEK million of trade gains
achieved for that country. (SEK 1=USD .12) The payoff factors for countries #1-#11
were SEK 3, 4, 2, 15, 7, 8, 8, 15, 1.5, 1 and 10, respectively. The expected incentive
payment per subject and period, calculated on the basis that the outcome would not
deviate much from the competitive outcome, ranged from SEK 70 to SEK 130, in
addition to a fixed payment of SEK 25 per period (over and above the show-up fee of
SEK 200).16 Any net losses at the end of the period would be deducted from this fixed
payment, up to a maximum deduction of SEK 25. Hence, although subjects could not
lose money by participating in the experiment they could “lose” money by engage in
trade, and it would be possible for them to gain considerable amounts of money even
in case they would make initial losses, not exceeding the fixed payment by too much.
Subjects’ experimental earnings came to range from SEK 3 to SEK 2 075 (net of the
show-up fee and income taxes).
Each subject participated in two trading periods (each 35 minutes long) playing the
role of the same country in both periods. In order to avoid repeated game effects, the
two periods used different demand and supply conditions and subjects did not receive
any information about the relevant situation for the second period until after the
completion of the first period. The environment of the second period deviated only a
16
The likelihood that subjects would “play” against the experimenter by allocating a trade surplus to
the subject(s) with the highest payoff factor(s) was deemed low since (a) each subject was anonymous
13
little from that used in the first, whereby it could be regarded as mimicking the end of a
subsequent five year emission budget period.
4.
The Results
Benchmark cases
Given the hypothesis stated in Section 2, the natural benchmark is the competitive
(efficient) outcome. Market efficiency is defined as the achieved share of the maximum
cost reduction (trade gains) of emission quota trading, which equals the difference in
aggregate costs between the case in which all countries fulfill their commmitments
unilaterally and the case of fully efficient emission quota trade. Tables 1 and 2 show the
competitive outcome for the environments used in period 1 and period 2, respectively.
Both tables show
n the amount by which each country would need to reduce its emissions if the
emission reduction was made unilaterally (column 2),
n the costs of these reductions (column 3),
n the emission reductions in the efficient-trade case (column 4),
n the cost of these reductions (column 5) and
n the resulting trade gains (column 6).
A country’s efficient net export of emission quota units is given by the difference
between columns 4 and 2. The trade gains of an exporting country are calculated as
export revenues minus the additional abatement costs caused by export activities. For
an importing country, the trade gains equal the abatement cost savings minus the
expenditure on imported quota units.
to others during the trade, only the country code was revealed to others, and (b) the subjects were not
allowed to communicate with each other during the experiment.
14
Table 1 The perfectly competitive outcome in period 1, Mton and MSEK respectively
p* = MSEK 1 450/Mton
Unilateral
Emission reduction ex post trade:
Emission
Cost
Mton Cost
Net gain
reduction
Belgium (#0)
Germany (#1)
Greece (#2)
Spain (#3)
France (#4)
Italy (#5)
The Netherlands (#6)
Portugal (#7)
Sweden (#8)
The UK (#9)
The US (#10)
Japan (#11)
12
220
0
2 200
66 000
0
0
12 695.3
32 000
11 250
900
18 000
19 333.3
1 477 045.3
270 800
34
285
21
60
35
110
46
18
9
240
1 550
360
20 949
127 625
10 325
35 250
25 030
65 750
27 350
9 875
6 525
173 960
871 250
205 480
0
25
80
30
6
15
80
1 900
400
Total
2 768
13 150
32 625
20 125
55 750
2 270.3
9 750
7 100
8 425
2 775
77 333.3
105 795.3
7 300
1 910 223.9
2 768
1 567 825 342 398.9
Table 2 The perfectly competitive outcome in period 2, Mton and MSEK respectively
p* = MSEK 1 200/Mton
Unilateral
Emission reduction ex post trade:
Emission
Cost
Mton Cost
Net gain
reduction
Belgium
Germany
Greece
Spain
France
Italy
The Netherlands
Portugal
Sweden
The UK
The US
Japan
10
210
0
0
25
70
25
10
15
85
1 950
380
1 562.5
63 000
0
0
9 838
28 000
8 928.6
2 083.3
17 250
18 062.5
1 352 500
223 250
33
275
20
53
36
105
40
20
8
240
1 600
350
18 450
119 500
9 800
29 000
20 700
63 000
23 000
9 800
4 800
144 000
810 000
182 000
10 712.5
21 500
14 200
34 600
2 338
7 000
3 928.6
4 283.3
4 050
60 062.5
122 500
5 250
Total
2 780
1 724 474.9
2 780
1 434 050 290 424.9
The efficient net trade in period 1 and 2 reallocates 396 Mton and 387 Mton,
respectively, or 14% of the aggregate emission reduction. This (efficient) trade reduces
the cost of reaching the aggregate emission target of 2 768 and 2 780 Mton carbon by
18% and 17%, respectively for the two periods. The competitive outcome gives 34%
and 45% of the surplus to the buyers, respectively for the two periods. (Thus, as
mentioned above, trade based on private information regarding buyers’ and sellers’
15
underlying demand/supply schedules would be expected to give a price path that starts
out below and converges to the efficient price level.)
The potential cost savings of carbon emission trading (in terms of percent) may
seem quite small. However, two circumstances that tend to reduce the potential costs
savings should be observed. One is that the experiment as designed here mimics a
situation in which only a subset of OECD countries engage in emission quota trade.
Ceteris paribus, potential cost savings of emission trading are likely to increase with
the number of participating countries. Particularly so if countries equipped with large
amounts of low costs abatement options, such as Russia, the Ukraine and other
economies in transition as well as developing countries, are added to the set of
countries engaging in emissions trading. Another circumstance reflects the fact that the
initial quota allocation utilized here is the result of two political bargaining processes –
the one that led to the Kyoto protocol and the one that resulted in EU’s internal
burden sharing rule – which both reduce trade as well as the gains from trade.
According to available estimates of the MACs of the countries’ included here, these
processes have to some, possibly considerable, extent allocated larger emission quotas
to countries with higher MACs and vice versa.
Other benchmarks of possible interest are cases in which major buyers or sellers
collude to maximize their profits. Just to indicate the potential effect on the outcome of
such a coalition, we focus on the case where the two major buyers, US and Japan, act
jointly as a single profit maximizing monopsonist while the other countries behave as
price takers. The market in this coalition case (in accordance with standard economic
theory regarding market power) is assumed to be governed by an institution that
requires all trade to be conducted simultaneously at a uniform price. Then, US and
Japan maximize their joint gains by withholding demand in the amount of 153 Mton
and 129 Mton, or by 39% and 34% of the quantity demanded at the competitive
market in period 1 and 2, respectively. The monospony price equals 1 050 MSEK and
966 MSEK, or 72% and 80% of the efficient price level in period 1 and 2, respectively.
The joint profits of Japan and US amount to 188 090 MSEK and 169 708 MSEK
(166% and 133% of competitive outcome) in period 1 and 2, respectively. Aggregate
trade gains of the other countries amount to 65 871 MSEK and 56 689 MSEK (29%
and 35% of competitive outcome) in period 1 and 2, respectively. The efficiency of this
monospony case is as low as 74% (period 1) and 78% (period 2).
16
Experimental data
The results of the five sessions conducted are presented below. As mentioned above,
each session consists of two periods. However, period 2 of session one and period 1 of
session two suffered from problems not likely to arise in the case of real emission
trading.17 Therefore, these trading periods have been omitted from the analysis. The
relatively small number of sessions conducted, producing only five independent pairs of
observations, is explained by budget reasons of two kinds. First, the number of suitable
subjects were limited. Second, the approach of testing emissions trading using a
realistic experimental environment is expensive. The experimental cost of each session
amounts to approximately USD 900.
Table 3 presents a summary of the experimental outcome. The outcome of the
laboratory test is characterized by high efficiency. In six of eight periods more than
95% of potential (maximum) trade gains were realized. Subjects representing net
buying countries, and in particular those representing the dominant buyer country (the
US) were able to attain trade gains larger than a perfectly competitive market would
have allowed them to, especially in trading period 1. The rest of this section describes
the outcome of each session in detail.
Table 3 Efficiency rates and profit shares for expected net buyers, percent
Competitive outcome
1
2
Trading period
1.1
2.2
3.1
3.2
4.1
4.2
5.1 5.2
Efficiency
100
100
87
98
99
99
78
96
99
100
Profit shares
US
Japan, US and Sweden
31
34
42
45
56
50
64
79
45
49
44
91
107
100
66
68
59
68
43
47
17
During period 2, session one, software problems caused the MUDA program to duplicate
automatically each transaction. For this reason the ex post trade quota allocation differed substantially
from the efficient one, i.e., low efficiency was obtained. However, it can be noted that the prices were
close to the efficient level, the average price equals 98% of the efficient level. In period 1 of session
two, the subject representing Germany misunderstood the cost-schedule in a way that made him sell
far more quota units than was profitable for his country. Therefore, prices came to differ significantly
from the expected level, the average price being 53% of the efficient level. Consequently, subjects who
speculated in that prices eventually would approach the efficient level made losses. It is likely that this
outcome influenced the trading behavior in period 2 by making subjects skeptical to the information
given to them about the expected (outcome) price level in period 2.
17
Session 1
Table 4 presents the outcome in period 1, session one. The efficiency achieved equals
87.2%. Efficient net trade not carried out (defined as the sum of each country’s
absolute deviation from the efficient trade volume divided by two) amounts to 79
Mton, or 20% of the efficient net trade volume. Figure 2, showing all bids, asks and
contracts made during period 1, reveals that prices did not converge to the efficient
level. Prices at the end of the trading period were 93% of the efficient level. The
average price amounted to 80% of the efficient price level. Thus, in general, net buying
countries gained more and net selling countries less than had the market been perfectly
competitive. As much as 60% of realized trade gains ended up in the hands of US and
Japan.
Table 4 Outcome Period 1, Session 1, Mton and MSEK, respectively
p* = MSEK 1 450/Mton
Unilateral
Emission reduction ex post trade:
Emission
Cost
Mton Cost
Net gain
reduction
Belgium
Germany
Greece
Spain
France
Italy
The Netherlands
Portugal
Sweden
The UK
The US
Japan
12
220
0
0
25
80
30
6
15
80
1 900
400
2 200
66 000
0
0
12 695.3
32 000
11 250
900
18 000
19 333.3
1 477 045.3
270 800
32
292
20
60
33
107
45
17
38
187
1 592
345
18 162.5
138 346.7
8 942.9
35 250
22 125
61 497.5
26 875
8 537.5
75 400
105 635.5
926 173.3
184 766.2
Total
2 768
1 910 223.9
2 768
1 611 712 298 512
/Figure
9 131.5
19 950.3
17 687.1
42 285
565.3
1 802.5
5 975
4 440.5
-26 655
46 080.8
166 730
10 518.8
2 about here/
Session 2
In period 2, session two, efficiency achieved equaled 98.5%, see Table 5. Efficient net
trade not carried out amounts to 13 Mton, or 3% of the efficient net trade volume.
Figure 3 shows that the price level converged to the efficient price from below,
average price being 71% of the efficient level. Hence, again, net buying countries
gained more and net selling countries less than they would had the market been
18
perfectly competitive. Almost 80% of realized trade gains ended up in the hands of net
buying countries – Japan, US and Sweden – as compared to 45% in the perfectly
competitive benchmark case. However, as mentioned, the mistake in period 1 made by
the subject representing Germany (rendering a mean price of only 50% of the efficient
price level, see footnote 17) is likely to have produced a set of (price) expectations
quite different from the intended situation of common expectations. This circumstance
may explain why speculation did not force prices to approach the efficient price level
faster than they did. Therefore, it may not be straightforward to interpret the outcome
in period 2 as the result of emission trade among well informed traders having common
expectations regarding the efficient price level.18
Table 5 Outcome period 2 of session 2, Mton and MSEK, respectively
p* = MSEK 1 200/Mton
Unilateral
Emission reduction ex post trade:
Emission
Cost
Mton Cost
Net gain
reduction
Belgium
Germany
Greece
Spain
France
Italy
The Netherlands
Portugal
Sweden
The UK
The US
Japan
10
210
0
0
25
70
25
10
15
85
1 950
380
1 562.5
63 000
0
0
9 838
28 000
8 928.6
2 083.3
17 250
18 062.5
1 352 500
223 250
33
286
20
53
36
105
40
10
7
238
1 602
350
18 450
133 783.3
9 800
29 000
20 700
63 000
22 999.6
2 083.3
3 450
140 845
812 404
182 000
3 897.5
6 760.7
-429
18 782
-1 439
1 292
16 261
3 360
7 679
11 971.5
183 467
34 357
Total
2 780
1 724 474.9
2 780
1 438 515 285 960
/Figure 3 about here/
18
As mentioned in footnote 11, it has been observed that initial prices, when trade is based on only
private information, tend to be such that trade partners divide the available surplus equally, Davis and
Holt (1993). The first contract in period 2 was concluded between Japan and the UK at a price of
MSEK 1 000 per Mton, very close to the price of an equal split (MSEK 987 per Mton). The next five
contracts were concluded between countries the Netherlands and the UK at a price level of MSEK 700
per Mton, again closer to the price level of an equal split (MSEK 569 per Mton) than to the efficient
price level of MSEK 1 200 per Mton. Hence, the outcome observed here resembles the outcome that
can be expected when trade is conducted upon private information regarding individual demand and
supply functions.
19
Session 3
The outcomes in period 1 and 2, session three, are shown in Table 6 and Table 7,
respectively. The efficiency achieved in the two periods equals 99% and 99.1%,
respectively. Efficient net trade not carried out amounts to 25 Mton and 31 Mton, or
6.3% and 8% of the efficient net trade volume, respectively. In both periods, marginal
prices were close to the efficient levels, see Figures 4 and 5. However, in period 1
early prices were below the efficient level, while period 2 exhibits a flat price path.
Average prices in the two periods equal 90.5% and 85% of the efficient levels,
respectively. The outcome in period 1 is quite near the competitive outcome, although
the dominant buyer (the U.S.) gained more (144% of the competitive outcome).
Aggregate consumer surplus amounts to 49% of realized gains as compared to 34%
under perfect competition. In period 2, as much as 91% of realized trade gains went to
the net buying countries (Japan, US and Sweden). However, since this outcome is due
mainly to the performance of the subject representing Japan who successfully
conducted a large number speculative transactions it can not be taken as an indication
of the US being able to exert market power (US’s trade gains exceeds only slightly
what it would have received under perfect competition).
Table 6 Outcome period 1 in session 3, Mton and MSEK, respectively
p* = MSEK 1 450/Mton
Unilateral
Emission reduction ex post trade:
Emission
Cost
Mton Cost
Net gain
reduction
Belgium
Germany
Greece
Spain
France
Italy
The Netherlands
Portugal
Sweden
The UK
The US
Japan
12
220
0
0
25
80
30
6
15
80
1 900
400
2 200
66 000
0
0
12 695.3
32 000
11 250
900
18 000
19 333.3
1 477 045.3
270 800
39
283
20
61
35
112
40
18
8
225
1 563
364
29 244.4
124 750
8 943.5
32 732.5
24 925
68 358.6
20 000
9 875
5 089.2
152 929.6
882 746.1
211 607
9 905.6
22 149
10 678.5
27 920.5
7 365.3
6 508.4
3 610
11 949
8 160.8
72 484.7
151 250.2
7 041
Total
2 768
1 910 223.9
2 768
1 571 201 339 023
20
Table 7 Outcome period 2 in session 3, Mton and MSEK, respectively
p* = MSEK 1 200/Mton
Unilateral
Emission reduction ex post trade:
Emission
Cost
Mton Cost
Net gain
reduction
Belgium
Germany
Greece
Spain
France
Italy
The Netherlands
Portugal
Sweden
The UK
The US
Japan
10
210
0
0
25
70
25
10
15
85
1 950
380
1 562.5
63 000
0
0
9 838
28 000
8 928.6
2 083.3
17 250
18 062.5
1 352 500
223 250
33
274
20
52
36
100
33
20
8
227
1 631
346
18 450
118 272.7
9 800
27 815.4
20 700
57 142.9
15 557.2
9 800.3
4 800
128 822.5
848 161
177 250
-5 532.5
38 825.3
13 472
30 749.6
-5 211
6 636.1
-4 118.6
20 593
4 050
-69 136
127 228
130 347
Total
2 780
1 724 474.9
2 780
1 436 572 287 902.9
/Figure 4 and Figure 5 about here/
Session 419
Table 8 and Table 9 show the outcomes in period 1 and 2, session 4, respectively. The
efficiency rates equal 78.1% and 95.6%, respectively. Efficient trade not carried out
amounts to 31% and 12% in period 1 and 2, respectively. In period 1, the price path
started from a low level and increased quite rapidly (Figure 6). For a while prices were
above the efficient level. However, at the end of the trading period prices fell to a level
somewhat below the efficient one. In period 2, prices were quite close to the efficient
level from start (Figure 7). Average prices amounted to 66% and 92% of the efficient
price in period 1 and 2, respectively, while prices at the end of the trading period were
86% and 90-100% of the efficient levels. As much as 107% of realized trade gains
were allocated to the US, an outcome partly explained by the losses made by Germany
and the Netherlands, which sold quota units at prices far below the efficient price, and
Portugal and Japan, which sold more units than profitable for them. In period 2 the US
received 66% of the trade gains realized.
19
The original session 4 has been replaced by a new session. The reason is that the subject vital for
the experiment (the US) had not studied the common information given to the subjects prior to the
21
Table 8 Outcome period 1 in session 4, Mton and MSEK, respectively
p* = MSEK 1 450/Mton
Unilateral
Emission reduction ex post trade:
Emission
Cost
Mton Cost
Net gain
reduction
Belgium
Germany
Greece
Spain
France
Italy
The Netherlands
Portugal
Sweden
The UK
The US
Japan
12
220
0
0
25
80
30
6
15
80
1 900
400
2 200
66 000
0
0
12 695.3
32 000
11 250
900
18 000
19 333.3
1 477 045.3
270 800
31
267
19
53
39
101
46
39
6
210
1 496
461
16 853.1
105 245
7 696.4
22 692.5
31 125
53 577.5
27 350
52 987.5
2 899.8
133 218.7
787 703.3
401 297.3
2 191.9
-17 400
10 183.6
7 607.5
-844.7
5 807.5
-14 330
-17 697.5
6 140.2
24 849.6
285 381
-24 311.3
Total
2 768
1 910 223.9
2 768
1 642 646 267 578
Table 9 Outcome period 2 in session 4, Mton and MSEK, respectively
p* = MSEK 1 200/Mton
Unilateral
Emission reduction ex post trade:
Emission
Cost
Mton Cost
Net gain
reduction
Belgium
Germany
Greece
Spain
France
Italy
The Netherlands
Portugal
Sweden
The UK
The US
Japan
10
210
0
0
25
70
25
10
15
85
1 950
380
1 562.5
63 000
0
0
9 838
28 000
8 928.6
2 083.3
17 250
18 062.5
1 352 500
223 250
33
270
21
55
21
110
40
22
8
214
1 607
379
18 450
113 666.7
11 100
31 760
6 941.69
69 500
23 000
13 100
4 800
114 490
818 449
221 705.8
4 462.5
22 848.3
17 700
32 956
-2 319.7
6 573
4 068.6
4 183.3
3 755
-2 821.5
183 459
2 647.2
Total
2 780
1 724 474.9
2 780
1 446 963 277 512
/Figure 6 and 7 about here/
Session 5
The outcomes in period 1 and 2, session 5, are shown in Table 10 and Table 11,
respectively. The efficiency rates equal 99.1% and 99.8%, respectively. Efficient trade
not carried out amounts 15 Mton and 15 Mton, or 3.8% and 3.9% of efficient trade
experiment, something she confessed after the session. Thus, the outcome of this session could not be
22
volume, respectively. The price level converged to the efficient level in both periods
(see Figures 8 and 9). However, while the price path started at a low level in period 1,
giving an average price equal to 74% of the efficient level, a relatively flat price path
could be observed in period 2 (the average price being 98% of the efficient price). The
share of trade gains going to net buying countries amounted to 68% and 47% in period
1 and 2, respectively. This discrepancy is mainly due to the behavior of the subject
representing the UK, who made substantial losses in the first period and a quite large
profit in the second.
Table 10 Outcome period 1 in session 5, Mton and MSEK, respectively
p* = MSEK 1 450/Mton
Unilateral
Emission reduction ex post trade:
Emission
Cost
Mton Cost
Net gain
reduction
Belgium
Germany
Greece
Spain
France
Italy
The Netherlands
Portugal
Sweden
The UK
The US
Japan
12
220
0
0
25
80
30
6
15
80
1 900
400
2 200
66 000
0
0
12 695.3
32 000
11 250
900
18 000
19 333.3
1 477 045.3
270 800
31
286
21
60
25
108
48
18
9
244
1 556
362
Total
2 768
1 910 223.9
2 768
16 853.1
129 086.7
10 325
31 250
12 695.3
62 893.3
30 675
9 875
6 525
179 910
872 501.8
208 405.5
13 046.9
111 280.3
8 835
20 200
1 019
4 856.7
5 372
495
7 865
-55 712.7
200 227.5
21 743.5
1 570 996 339 228
Table 11 Outcome period 2 in session 5, Mton and MSEK, respectively
p* = MSEK 1 200/Mton
Unilateral
Emission reduction ex post trade:
Emission
Cost
Mton Cost
Net gain
reduction
Belgium
Germany
Greece
Spain
France
Italy
The Netherlands
Portugal
Sweden
The UK
The US
10
210
0
0
25
70
25
10
15
85
1 950
1 562.5
63 000
0
0
9 838
28 000
8 928.6
2 083.3
17 250
18 062.5
1 352 500
interpreted in any meaningful way.
33
274
20
53
36
104
40
20
8
240
1 615
18 450
118 306.7
9 800
29 000
20 700
61 805.7
23 000
9 800
4 800
144 000
828 225
10 087.5
32 213.3
11 761
27 820
2 302
5 071.3
2 988.6
2 863.3
5 495
58 172.5
123 333
23
Japan
380
223 250
337
166 882.9 7 597.1
Total
2 780
1 724 474.9
2 780
1 434 770 289 705
/Figure 8 and Figure 9 about here/
5.
Conclusions
The likelihood that at least some countries would engage in intergovernmental
emissions trading instead of letting their firms trade internationally has raised the
question whether large countries would be able to exert market power. The laboratory
tests reported here investigate a case of international carbon emissions trading (during
the crucial last six months of the five-year-long trading period, as indicated in the
Kyoto protocol) with one country (the US) assumed to be responsible for about 90
percent of the efficient net trade volume. According to standard economic theory, such
a situation would lead to an inefficient outcome. However, this theory does not take
into account that the outcome may depend on the rules governing trade. In Section 2 it
is hypothesized that if trade is governed by so-called double-action rules, then (i) the
presence of large traders would not hinder trade from being efficient and (ii) dominant
traders would not be able to influence prices to their advantage during an end period
characterized by nearly full information regarding the underlying net-demand.
The main findings of the experiment are that
1. the presence of large countries is not likely to stop trade from being highly
efficient and
2. the dominant trader seems to have difficulties to maintain an influence over
prices after a first trading period.
In six of eight periods more than 95 percent of potential (maximum) trade gains
were realized (see Table 3). The two trading periods each session consisted of exhibit
no systematic differences regarding the efficiency rate. In other words, we do not
observe that the dominant buyer country exerts market power by withholding demand
from the market as predicted by standard economic theory. Although the number of
observations is to small for applying statistical tests, these results seem to suggest that
24
the presence of large countries on a carbon emission quota market organized as a
continuous double auction is unlikely to stop trade from being highly efficient.
Four of the eight periods exhibit a flat price path quite near the efficient price level,
while prices converged to the efficient level from below in the other four. So, in one
half of the periods the outcome is consistent with hypothesis (ii). The outcome in the
other half is, however, similar to what can be expected when (a) a dominant trader has
been successful in utilizing the scope for price discrimination that the double-auction
rules allow or (b) trade is based on private information only, i.e., the “convergence
bias” phenomenon discussed in Section 2.
It is interesting to note that three of the four “flat-price-path cases” occurred in the
second trading period. This suggests a hypothesis that even if the dominant trader is
able to exert market power in the first period, in a way that does not reduce market
efficiency, such behavior can not be maintained in the second trading period.
Another, perhaps even more likely, explanation to the difference between the first
and second periods is of a methodological kind. In contrast to the often used approach
of letting subjects gain experience by trade repeatedly, this experiment used extensive
written information and training rounds with demand and supply conditions different
from those of the test case in order to have experienced and well informed subjects. It
is possible that the subjects, in spite of these efforts, needed more time and/or an
opportunity to trade at least once under conditions similar to those used in the test in
order to acquire the capacity of qualified traders. To distinguish between this
hypothesis from the one in the preceding paragraph and to confront the extent of
hypothesis (ii) behavior over several periods further testing is needed.
25
References
Bohm, Peter, 1995, “An Analytical Approach to Evaluating the National Net Costs of a
Global System of Tradable Carbon Emission Entitlements”, United Nations,
Geneva.
Bohm, Peter, 1997, “Joint Implementation as Emission Quota Trade: An Experiment
Among Four Nordic Countries”, Nordic Council of Ministers.
Bohm, Peter and Björn Carlén, 1999, “Emission quota trade among the few: laboratory
evidence of joint implementation among committed countries”, Resource and
Energy Economics, vol. 21 nr. 1.
Bohm, Peter, 1998, “Determinants of the benefits of international carbon emissions
trading: theory and experimental evidence”, in: Emissions Trading - Proceedings
of the Conference on Greenhouse Gas Emissisions Trading (Sydney May 21-22,
1998), ABARE, Canberra.
Capros, P., T. Kotsomiti S. Georgakopoulos and Filippoupolitis A., 1997, “Macroeconomic Implications of the “Kyoto” CO2 Target for the EU” - Report to
European Commission DG-XII under Climate Technology Strategy JOULE
Project, National Technical University of Athens.
EU, 1996, “Quantified Emission Limitation and Reduction Objectives within Specified
Time Frames”, Ad Hoc Group on Climate Italian Presidency, Rome June 8,
1996.
Easley, David and John O. Ledyard, 1993, “Theories of Price Formation and Exchange
in Double Oral Auctions”, in The Double Auction Market – Institutions,
Theories and Evidence, Santa Fe Institute.
Friedman, Daniel, 1991, “A Simple Testable Model of Double Auction Markets”,
Journal of Economic Behavior and Organization, vol. 15, 1991.
Friedman, Daniel, 1993, “The Double Auction Market Institution: A Survey”, in The
Double Auction Market – Institutions, Theories and Evidence, Santa Fe
Institute.
Gjerdstad, Steven and John Dickhaut, 1995, “Price Formation in Double Auctions”,
Discussion Paper No. 284, Department of Economics, University of Minnesota,
November 1995.
Godby, Robert W., 1997, The Effects of Market Power in Emission Permit Markets,
Ph.D. thesis, McMaster University, 1997.
Godby, Robert W., Stuart Mestelman and R. Andrew Muller, 1998, “Experimental
Tests of Market Power in Emission Trading Markets”, Working paper
Department of Economics, McMaster University, 1998.
26
Hahn, Robert, W., 1984, “Market Power and Transferable Property Rights”, The
Quarterly Journal of Economics, November 1984.
Hizen, Yoichi and Tatsuyoshi Saijo, “Designing GHG Emissions Trading Institutions in
the Kyoto Protocol: An Experimental Approach”, forthcoming in Environmental
Modeling and Software.
Holt, Charles, A., 1995, “Industrial Organization: A Survey of Laboratory Research”, in
The Handbook of Experimental Economics, John H. Kagel and Alvin E. Roth,
editors.
IPCC, 1996, Climate Change 1995 - Economic and Social Dimensions of Climate
Change, Cambridge University Press.
Ledyard, John O. and Kristin Szakaly-Moore, 1994, “Designing Organizations for
Trading Pollution Rights”, Journal of Economic Behavior and Organization,
vol. 25, 1994.
Misiolek W. S. and H. W. Elder, 1989, “Exclusionary Manupilation of Market for
Pollution Rights”, Journal of Environmental Economics and Management, vol.
16, 1989.
Montgomery, D. W., 1972, “Markets in Licenses and Efficient Pollution Control
Journal of Economic Theory.
Muller, R. Andrew, Stuart Mestelman, John Spraggon and Robert W. Godby, 1999,
“Can Auctions Control Market Power in Emissions Trading Markets?”, mimeo,
1999.
Plott, Charles, R., 1989, “An Updated Review of Industrial Organization: Applications
Handbook of Industrial Organization, vol. II, R.
Schmalensee and R. D. Willig, editors, Amsterdam: North Holland, 1109-76.
Sandor, Richard L., Joseph B. Cole and Eileen M. Kelly, 1994, “Model Rules and
Regulations for a Global CO2 emissions Credit Market”, in: Combating Global
Warming, United Nations, New York, 1994.
Smith, Vernon, L., and Arlington W. Williams, 1982, “The Effects of Rent Asymmetries
in Experimental Auction Markets”, Journal of Economic Behavior and
Organization, 3, 99-116.
Tietenberg, Tom, H., 1985, Emissions Trading - an exercise in reforming pollution
policy, Resource for the future Inc.
United Nations, 1997, “Kyoto Protocol to the United Nations Framework Convention
Westskog, Hege, 1996, “Market Power in a System of Tradable CO2 Quotas”, The
Energy Journal, vol. 17, No. 3.
27
Appendix A Instructions
The subjects received written information on two occasions. Two days before the
experiment, subjects were given written information common for all subjects. At the
beginning of the experiment, each subject received private information about which
country he/she was assigned to, the updated abatement cost schedule of that country,
and his/her payoff factor. This final section reproduces the information common to the
subjects and, for illustration, the private information given to the subject representing
country #0. Text in bold within parenthesis serves as explanations to the reader of this
paper.
Common Information (translated from Swedish)
1
Background
You will participate in an experiment that seeks to shed some light on the possibilities
to by the means of so-called emission trading reduce the costs of a climate treaty with
the objective of preventing significant changes in the global climate.
Background: The global community faces the following threat. The growing
concentration of so-called greenhouse gases (GHGs) in the atmosphere runs the
risk of leading to an increase in the average global temperature.20 Such a climate
change, if large, may give rise to rather drastic consequences, such as changes in
the wind and sea currents, reduced food production, higher sea level, and a
desertification. The conditions for localisation of settlements, farming and
industries would thereby be affected, resulting in possibly large population
movements. The international community has taken this threat seriously. At the
UN’s conference on environment and development in 1992 in Rio de Janeiro, the
so-called Climate Convention was adopted. The objective of the Climate
Convention is to stabilise the concentration of GHGs in the atmosphere at such a
level and within such a time frame that the above-mentioned changes will not be
drastic and that the ecological systems can adjust more slowly.
The single most important GHG is carbon dioxide (CO2). CO2 is emitted to
the atmosphere mainly through combustion of fossil fuels. There exists no
economically feasible technique to separate CO2 from other emissions of fossil
fuel combustion. Thus, in order to prevent climate changes the use of fossil fuels
has to be reduced. A specific feature of GHGs is that their effect on the global
climate does not depend, at least not to any significant extent, on where the
emission source is located.
Over 160 countries have signed the Climate Convention. In December last year
(1997), these countries gathered in Kyoto, Japan, to negotiate over country specific
emission targets for CO2 for the period after the year 2000. The Kyoto-meeting
resulted in a proposal to a climate treaty comprising a set of quantitative emission
targets for the industrialised countries. If a sufficiently large number of countries ratify
the proposal, the treaty will come into force. In what follows it is assumed that this is
20
GHG are gases that restrict the radiation of heat from the earth.
28
the case and that the Kyoto-protocol has entered into force. The emission quotas
(targets) for the largest OECD-countries are presented in Table 1 below.
The costs of reducing the CO2 emissions differ between countries. A given
aggregated emission target for a group of countries is attained at lowest possible cost
when each country has reduced its emissions by the amount that yields equality in
marginal abatement costs for all countries. Lower emissions of CO2 presupposes a
limitation of the use of fossil fuels, which dominates the majority of the industrialised
countries energy system, whereby abatements of CO2 are associated with large costs
for these economies. Cost-effectiveness, i.e., to minimise these costs, is therefore of
great importance for international climate change policy.
The ambition of the Kyoto-meeting was to find an allocation of national emission
quotas that as many countries as possible could accept. In order to facilitate a costeffective allocation of the aggregated emission reduction the Kyoto-protocol allows
countries that are committed to binding emission targets to engage in emission trading
among themselves. Countries with high marginal abatement costs can on such a market
pay countries with lower costs to reduce their emissions by more than what their
national emission quotas imply so as to make all involved countries gain from this
trade. Trade with emission quotas does not differ in any important aspect from trade
with other goods and services. On a competitive market an equilibrium price would be
established so that the marginal abatement costs in different countries are equalised,
i.e., the cost-effective allocation of emission reductions is attained.
The objective of this experiment is to study the performance of trade with CO2
emission quota units when trade is conducted on an stock-exchange kind of market
and when the countries engaged in trading are of different sizes.
2 The experiment
The experiment intends to mimic the following assumed situation. A group of 12
countries committed to the national emission targets allotted to them by the Kyotomeeting (see Table 1) has decided to use the possibility that the Kyoto-protocol gives
to emission trading and to establish an organised market for emission quota units. With
this possibility to trade each country has to chose between
(a) reducing the national emissions by more than necessarily and sell the surplus of
emission quota units to some other country(-ies),
(b) increasing its own emission quota (target) buy paying some other country(-ies)
to reduce their emissions, or
(c) attaining its emission target only by reducing its own emissions.
Trade with emission quota units will be governed by so-called double-auction rules.
(The same rules that govern the majority of the large stock exchanges and markets for
primary goods throughout the world.) In the experiment the market will be electronic,
something that has been more and more common among the large stock markets. Each
country is connected to the market through a computer terminal. When the market is
open a country that wants to buy can via its computer send a buy order (unit price and
number of units) to a central computer. A country that wants to sell can in the same
manner send an sell order (unit price and number of units). The buy order with the
highest price and the sell order with the lowest price are made public as the standing
bids. This information is shown on each country’s computer monitor. The standing buy
29
order (sell order) is replaced when the central computer receives a higher (lower) buy
order (sell order). This rule tends to give adjustments of declining sell orders and
ascending buy orders until some agent accepts to trade to the standing sell/buy price.
When a country accepts a standing buy (sell) order (or part thereof) this order is
removed (or reduced by the accepted volume). All acceptances of standing sell/buy
orders leads to binding transactions. This double-auction market does not use any
order book/queue. The implication of this is that when the volume of a standing
bid/ask has been accepted the next buy/sell order that the central computer receives
will be the standing bid. It is possible for a lower (higher) buy (sell) order to become
the standing bid. The price levels to which emission quota units are traded can
therefore increase and decline during the trading period.
On this market it is assumed that the trade unit is 1 Mton. That is, quantities
offered/asked can be values such as 7 Mton, 3 Mton and 1 Mton. Prices can only be
stated as whole MSEK/Mton.
Countries that will engage in real emission trading have strong incentives to gather
information about each others expected emission levels. Table 1 presents the countries’
expected CO2 emission levels in the year 2010 and their emission targets that year. The
emission reduction each country is responsible for amounts to the difference between
the expected emission level and the emission quota (target).
Table 1 Emission levels and emission reductions in the year 2010, millions ton CO2
Emissions
2010
Emission target
2010
Emission reduction
2010
Country #0
Country #1
Country #2
Country #3
Country #4
Country #5
Country #6
Country #7
Country #8
Country #9
Country #10
Country #11
112
970
90
250
395
460
170
63
80
610
6 500
1 400
100
750
90
250
370
380
140
57
65
530
4 600
1 000
12
220
0
0
25
80
30
6
15
80
1 900
400
Total
11 100
8 332
2 768
The countries would also have strong incentives to collect information about each
others’ costs of reducing emissions of CO2. It is unlikely that countries would be able
to successfully disguise from other countries large abatement opportunities and the
cost of these. Thus, it is here assumed that the countries have been so successful in
gathering information about each others’ costs measures that they have common
expectations about these cost schedules.
With the concept marginal abatement cost (MAC) is here meant the cost of
reducing the emissions of CO2 by one unit. MAC is the sacrifice (the value of the
reduction in consumption and production) a country needs to do in order to reduce its
emissions of CO2. In Appendix 1 the countries’ MACs are presented in diagrams.
Figure 1 below illustrates such a diagram for an assumed country.
30
Figure 1 illustrates country Xempel’s demand curve for CO2 emissions, equal to the
country’s marginal valuation (MV) of emissions. Without any climate treaty the
country would increase its emissions until its valuation of further emissions equals
zero, i.e., the country’s emission level would be where the MV-curve intersects the xaxis. This level is the forecasted emission level and is labeled business-as-usual (BAU)
level. Emission levels below the BAU level imply that incomes are sacrificed, i.e., costs
for the country. The cost of reducing emissions from the BAU level is indicated by the
MV-curve read from right to the left. The curve thus, illustrates the country’s marginal
costs of reducing CO2 emissions.
The vertical line in the diagram marked with target level (the target line) states the
emission level the country is committed to not to exceed. In case a country chooses to
accomplish its emission target unilaterally, i.e., only by domestic emission reduction,
the country’s MAC is denoted by the point where the MV-curve intersects the target
line. A country’s total cost of achieving its emission target unilaterally is given by the
area under the MV-curve between the target level and the BAU-level (the shaded
area).
In order to avoid talking about prices that can be perceived as related to the real
experimental situation we use in this and subsequent examples a hypothetical monetary
unit, MFrang.
Figure 1. Country Xmple’s Marginal valuation (marginal cost) of emissions,
MFrang/Mton CO2.
(For an illustration of this type of diagram the reader is referred to Figure 1 in
the main text.)
The point of departure for interpreting this type of diagram is that the country in
absence of trade will be at the point where the MV-curve intersects the target line. A
country makes trade gains by purchasing emission quota units, i.e., increasing its
emission quota, to prices lower than the country’s valuation of additional emissions
(marginal cost of reducing its emissions). Similarly, a country makes trade gains also
by selling emission quota units, i.e., reducing its emissions by more than required by
the emission target, to prices that are higher than the country’s marginal abatement
cost. Figure 1 shows that country Xmpel’s MV-curve and target level gives an
underlying demand/supply schedule such that the country can make trade gains by
selling quota units (go to the left from the target line) to prices above 300
MFrang/Mton and/or purchasing emission quota units (go to the right from the target
level) to prices below 300 MFrang/Mton. Hence, whether country Xmpel will act as a
seller or a buyer of emission quota units depends on the price levels that are established
on the market. It is possible that the price path is such that country Xmpel finds it
profitable to first act as a buyer and then as a seller, and vice versa.
As shown in Appendix 1, the expected MAC differ among the countries. These
differences reflect among other things differences in endowments of energy,
geographical and economical conditions. If each country chooses to accomplish its
emission target unilaterally, the MACs would vary between 0 MSEK/Mton (country
#2 and country #3) and 2 4000 MSEK/MTON (country #8). Hence, the aggregated
emission target can be attained at a lower cost if the countries engage in emission
quota trade.
31
Maximum potential trade gain
Given information about the countries’ expected MV-curves and emission targets
(Table 1) the countries’ expected underlying demand/supply schedules for emission
quotas can be calculated. (The individual as well as the aggregate underlying
demand/supply conditions are presented in Appendix 1.) Equipped with this
information it is possible to calculate the (expected) perfectly competitive outcome,
which gives information about the expected efficient price (P*), the countries’ net
demand and supply at this price level and an idea of how trade gains are distributed
among countries in the cases price levels on the market would not deviate much from
the expected efficient level. Under perfect competition the market price is such that the
countries’ MAC would be equalized. The expected efficient price would amount to 1
500 MSEK/Mton and the trade among the 12 countries to 395 Mton. The trade with
emission quota units and the associated surplus would be distributed as shown by
Table 2.
Table 2 Trade under perfect competition, Mton and MSEK, respectively
Unilateral
Emissions- Cost
reduction
P* = MSEK 1 500/Mton
Emission reduction ex post trade:
Mton Cost
Net gain:
#0
#1
#2
#3
#4
#5
#6
#7
#8
#9
#10
#11
12
220
0
0
25
80
30
6
15
80
1 900
400
2 000
66 000
0
0
13 750
32 000
11 250
900
17 250
20 000
1 502 727
275 000
36
285
22
59
34
110
45
17
10
240
1 550
360
Σ
2 768
1 940 877 2 768
% of unilateral
cost
SEK
per capita
22 500
131 250
11 500
33 300
25 450
66 500
26 250
9 500
7 500
180 000
877 500
207 000
15 500
32 250
21 500
55 200
1 800
10 500
7 500
15 600
3 000
80 000
100 227
8 000
775
49
13
33
67
878
13
400
7
3
1 532
396
2 062
1 410
32
184
488
1 574
342
1 370
385
64
1 598 250
342 627
18
467
The difference between column 2 and 4 states the countries’ expected net trade
under perfect competition. A negative value indicates that the country is a net seller
and a positive value that the country is a net buyer. In the case illustrated in Table 2
three countries appear as net buyers (countries #8, #10 and #11) and nine countries as
net sellers. Note that the buyer side is strongly concentrated with a single agent
(country #10) answering for almost 90% of the demand at the efficient price level.
As indicated earlier, countries have strong incentives to prior to the emission quota
trade gather information about each others underlying demand/supply conditions,
whereby the remaining uncertainty would be quite small. Before the trade starts on
Wednesday (9/12) you will receive an updated information about the underlying
demand/supply schedule (the true MV-curve) for the country you will represent.
32
The trading situation/traders’ information
It is not obvious that all, if any, trade will be conducted at P* so that the outcome
would be as Table 2 indicates. On a exchange as this one, different transactions can be
concluded at different prices. Moreover, no agent can be taken as a price taker, no
matter how small it is. Each country decides what price and quantity to bid (and at
which point in time to send it) and at which price levels he/she is willingly to accept to
buy/sell emission quota units. The prices at which trade will be conducted are
determined by the negotiation process that takes place on the market. At this market
all transactions are binding, i.e., they cannot be annulled. Revenues and expenditures of
agreements already concluded are thus not affected by additional trade. This means
that each agent has incentives to buy/sell additional quota units as long as the standing
sell/buy price (the price at which someone is willingly to sell/buy quota units) is
lower/higher than the agent’s MAC. This implies that this type of market tends to
generate buy/sell prices at the end of the market day that lie close to P*. It has been
shown that this is true also for the cases where initial buy/sell prices deviate
significantly from the efficient level.
In a real trading situation large values will be at stake. The countries can therefore
be expected to engage experienced and competent traders to represent them at the
market for emission quota units. As accounted for above, traders on this market can be
taken to being well informed about other countries’ expected underlying
demand/supply conditions and thereby having a pretty good perception about P*.
Experienced traders are familiar with the functioning of this type of markets and, in
particular, that it tends to generate end (marginal) prices close to the efficient level.
Such traders are also aware of the opportunity to profitable speculative trade that arise
if the price would deviate substantially from P*.
With profitable speculative trade is meant the following. If the standing buy price
during a period would be significantly higher than P* and later on converge to that
level, it would be possible to make trade gains by selling emission quota units during
the “high-price” period and buy them back later on when the price has declined.
Correspondingly, it would be possible to make profitable speculative trade if the
standing sell prices during a period were lower than P*. Experienced traders monitor
the market carefully in order to trade at best possible prices and, when the opportunity
arises, take advantage of price variations.
Hence, the scope of profitable speculative trade implies that a trader may have
incentives to buy/sell larger volumes than his/her country’s underlying demand/supply
schedule actually indicates. A trader that represent a country with a large underlying
net demand of quota units (at prices around P*) and that is convinced that prices at the
end of the trading period will be close to P* might act as follows. At sell prices above
P* he/she is withholding demand and may even try to sell quota units if the price
difference is sufficiently large, in order to buy emission quota units at lower prices later
on. At sell prices below P* the trader buys large volumes, even more than the
underlying demand/supply schedule actually permits, with the prospects of selling these
units at higher prices later on.
Correspondingly, a trader representing a country with a large underlying supply (at
the efficient price level) might act as follows: At buy prices below P* the agent
withholds supply and it might even buy quota units in order to sell these units at a
higher price later on. At buy prices that exceed the efficient level by much, the trader
33
might try to sell large volumes, with the ambition to buy them back later on when the
prices are lower.
To sum up: On a exchange of this kind, the price on emission quota units can vary
during the trading period. However, at the end of the trading period the market tends
to have generated sell/buy prices that lie close to the efficient price level. Real traders
can be taken to observe the market carefully in order to trade at best possible prices
and, if the opportunity arises, try to make speculation profits. The implications of this
is that the current demand (supply) may deviate from the sum of all countries’
underlying demand/supply schedules. With experienced and well informed traders it is
possible that the demand for quota units, when sell prices are below the efficient level,
may well exceed the underlying demand of the market. At instances when buy prices
are above the expected price, it is, correspondingly, possible that a supply far greater
than the underlying supply of the market is materialised.
Please observe! Even if, as has been emphasised here, all traders/countries have
strong incentives to collect information about other countries’ MV-curves, it has to be
stressed that no one with 100% certainty can know whether the available information
is correct or if this information actually will guide the trade. Moreover, it is also
possible, of course, that information about the conditions in other countries contains
certain approximations. Political aspects of potential abatement measures in a country
may lead to other decisions about such measures than those reflected by the calculated
MV-curves. Even if these uncertainties, due to the incentives for information
gathering, can be expected to be very small, is it possible that they affect the basis for
the trade of marginal units, i.e., when trade gains for involved parties according to the
given information are small. The implications of this is, nevertheless, a remaining
uncertainty about the efficient price level. This uncertainty is reflected on the market
that you will trade on in the way that the information about the MV-curves of the
countries that is presented in Appendix 1 may deviate somewhat from the true MVcurves. This means that each trader will, just before the market opens for trade, receive
an updated version of the MV-curve of the country he/she will represent.
3. Your Assignment
On Wednesday your will represent one of the above presented countries (country #0 11) on the market for emission quota units. Your assignment is to try to minimise your
country’s costs of achieving the emission target it is committed to through the Kyotoprotocol.
The point of departure is that the country has calculated the relevant costs for
unilaterally achieving its emission target and has given to its trader, you, a trading
budget that enables you to act on the market for emission quota units. If you
accomplishes the country’s emission reduction at a lower cost than the one of attaining
the target unilaterally by purchasing (selling) emission quota units at prices below
(above) your country’s MAC, i.e., making trade gains, you will be paid a certain
fraction of this trade gain. On Wednesday you will receive information about the size
of this fraction. At that time you will also be given private information about which
country (of the countries #0 - 11) you will represent and that country’s true underlying
demand/supply schedule. Your expected payment amounts to 400 - 500 SEK.
Below is presented, for the one that wishes to confirm that he/she has understood
the text above, a stylised example of how country Xmpel can act on the emission quota
34
market. As has been mentioned earlier, we use in our examples a hypothetical
monetary unit, MFrang, in order to avoid to talk about prices that can be perceived as
related to the real experiment situation.
Example
Assume that you represent country Xmpel, which is committed to an emission
reduction of 8 Mton and has MAC according to Figure 1. The costs of the country
achieving its emission target unilaterally amounts to 1 2000 MFrang. (The shaded
area.) A possible sequence of happenings is the following. At the beginning of the
trading period you announce a sell order, which becomes the standing sell order: 1
Mton at the price 500 MFrang/Mton. This bid, if accepted, would give your country a
( 400 − 300 ) MFrang
trade gain of 150 MFrang (= 500 MFrang x 1 Mton x 1 Mton
2
- 300 MFrang x 1 Mton, where the first term states revenues and the other two the
area beneath the MC-curve in the interval 16 to 17 Mton). Assume, however, that this
bid was not accepted. We assume that you, after a relative short period find out that
there exist countries willingly to sell quota units at prices around 200 MFrang/Mton,
a price level at which you are prepared to buy 2 Mton. You send a bid with the
content that you accept to buy 2 Mton at the price 200 MFrang per Mton, which is
registered before any other buy bids/acceptances whereby a transaction with that
content is conducted. This trade gives your country a trade gain of 100 MFrang (=
(300 − 200) MFrang
200 MFrang x 2 Mton +
x 2 Mton - 200 MFrang x 2 Mton,
2
where the two first terms states the abatement costs country Xmpel avoids by
conducting this trade and the third terms states the expenditures.). Thereafter you
find that the standing asks and bids lie around 200 MFrang, price levels which does
not allow you to make any larger trade gains. After a while the market has closed.
Your net trade is a purchase of 2 Mton for country Xmpels account, i.e., country
Xmpel is allowed to emit 19 Mton instead of 17 Mton.
Country Xmpel’s trade gains (= avoided abatement cost - net expenditures) is
according to this example 100 MFrang.
The result of this trade is, thus, that country Xmpel accomplishes the emission
reduction it is committed to at a lower cost than had the country achieved its emission
target unilaterally (i.e., if it would have conducted the whole emission reduction by
domestic actions). As mentioned above, your payment is proportional to the trade gain
of the country you represent. Hence, the payment you receive after the experiment
depends on how successful you are at the market for emission quota units.
It is not necessarily that you continuously calculate your country’s accumulated
trade gains to be able to act successfully on the market. It suffices to be able to
determine relatively fast whether or not your country will gain from conducting
different transactions. The important thing is to be able to calculate how your
country’s abatement costs change due to the transaction. Therefore, we ask you to
practice on how to calculate changes in the abatement costs.
One way to train on acting on this kind of market is that you by yourself play the
role of a couple of typical countries and contemplate how you would act given
different price paths.
35
4. Other things
On Wednesday you will, prior to the experiment, be given the opportunity to practice
on the type of electronic market that the emission quota trade will be based on. Before
the trade starts you will receive private information about which of the countries #0 #11 you will represent, that country’s true MV-curve as well as the fraction of the
trade gains you are allowed to keep. At that time a short presentation of instructions
will be held and any questions will be answered.
Private Information (Example of the information given to the subjects at the
beginning of the experiment.)
You have been appointed to negotiate for country #0. Your assignment is to minimise
your country’s cost for reaching the emission target the country by signing the Kyotoprotocol has committed itself to: That by the year 2010 not emit more than 100 Mton
CO2 (=your country’s initial inventory). The country’s updated MV-curve (see the
attached figure in which the largest changes are marked) indicates the BAU emission
level to 112 Mton CO2. Country #0 is thus committed to an emission reduction of the
amount of 12 Mton CO2.
The other countries’ emission targets, expected BAU emission levels as well as the
emission reductions they are expected to be committed to are presented in the common
information you have received (Table 1). There you also find information about the
other countries’ expected marginal abatement costs schedules.
The cost of your country to unilaterally attain its emission target amounts to 2 200
MSEK (the area beneath your country’s actual MV-curve – also it’s MC-curve – in the
interval the target level-the BAU level). By acting on the market for emission quota
units it may be possible for your country to make trade gains.
You have at your disposal an amount of 19 000 MSEK( = your country’s initial
cash on hand). You can use this amount to buy emission quota units from other
countries. Your are not allowed to buy emission quota units on credit. Hence, your
expenditures cannot exceed your cash on hand. If you sell emission quota units your
cash on hand increases with the sales revenues. Your cash on hand will automatically
be updated when you transact at the market.
You are not allowed to sell more emission quota units than the country has at its
disposal. Hence, your net sales of emission quota units can not, at any time, be larger
than 100 Mton (= your country’s initial inventory). Moreover, your country has only
calculated its marginal abatement costs over the interval 70-112 Mton CO2. This
means that your country does not allow your net sale at the end of the trading period
to imply an ex post trade emission quota that lies outside this interval. However,
nothing stops you from leaving this interval temporarily during the trading period.
Your payment consist of a fixed fee (200 SEK which you will receive after the
experiment) and an incentive payment. The incentive payment is calculated as follows:
Prior to any trade it amounts to 25 SEK. For each 1 000 MSEK in trade gains you
negotiate to your country the incentive payment increases with 6 SEK. (In case you
would make losses it will be reduced by 6 SEK per 1 000 MSEK loss. The incentive
payment can, however, never be lower than 0 SEK.)
36
Your country’s trade gains are calculated as described earlier.
Information given for the second period (Example of the information given to
the subjects at the beginning of the second trading period.)
1. Your assignment
Also this trading period your are representing country #0. Your assignment is as earlier
to try to minimise your country’s cost of reaching the emission target it is committed
to through the climate treaty: The emission target equals 100 Mton CO2 (= your
country’s initial inventory). Your country’s updated MV-curve indicates as the BAU
emission level 110 Mton CO2, see Appendix 2. Your country is thereby committed to
an emission reduction of 10 Mton CO2.
The cost for your country of attaining this emission target unilaterally amounts to 1
562.5 MSEK.
You have at your disposal an amount of 19 000 MSEK (=your country’s initial cash
on hand).
For every 1 000 MSEK in trade gains you negotiate for your country you earn 8
SEK. In all other aspects the rules are as in the first period.
As you can see, the trading situation differ somewhat from the previous period in
that the countries’ BAU emission levels, emission targets and expected underlying
demand/supply schedules are different. You can interpret this trading situation in the
following way. Those countries that the year 2010 were committed to binding emission
targets are taken to be committed also for subsequent periods. BAU and underlying
demand/supply schedules for a later year, say 2015, are likely to differ from the
conditions that prevailed the year 2010. It is also possible that the emission quotas for
the year 2015 would be different than those for the year 2010. The countries would in
such a sequential period utilise any lessons from earlier trading periods.
Below is given information about your country’s and the other countries’ emission
targets and forecasted emission levels (Table 1), expected underlying demand/supply
conditions (Appendix1) as well as the calculated perfectly competitive outcome (Table
2).
2. Common information
Table 1 Emission levels and emission reductions in the year 2010, millions ton CO2
(This Table is similar to Table 1 above)
In case each country chooses to attains its emission target unilaterally, the expected
MACs would vary between 0 MSEK/Mton (country #2 and country #3) and 2 400
MSEK/Mton (country #8).
The expected efficient price equals 1 200 MSEK/Mton and trade among the
countries equals 387 Mton CO2. This trade and the associated surplus would be
distributed as is presented in Table 2.
Table 2 Trade under perfect competition, Mton CO2 and MSEK, respectively
(Similar to Table 2 above)
37
The difference between column 2 and 4 states the countries’ net trade. A negative
value indicates that the country is a net seller of emission quota units and a positive
value that the country is a net buyer. In the case of perfect competition as illustrated in
Table 2, three countries act as net buyers (countries #8, 10 and 11) and nine countries
act as net sellers. Note that the buyer side is highly concentrated with a single buyer
(country #10) answering for about 90% of the demand at the expected efficient price.
Appendix 2.1 Expected MV-curves of all countries and the expected aggregated net
demand. (Not Shown here. See Appendix 1 for the same type of diagrams.)
Appendix 2: Country #0’s true MV-curve turned out to be the same as the expected
MV-curve presented in Appendix 2.1.
Appendix 1
(The diagrams show the expected MAC-curves of the twelve countries as well as
expected aggregate net demand. Emission levels in Mton CO2 are measured by
the x-axis and MSEK/Mton the y-axis. Vertical lines indicate the countries’
target levels.)
Country #0
Country #1
2000
1900
1800
1700
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
915
960
244
870
825
780
236
Country #2
735
690
645
600
116
109
95
102
88
81
74
67
60
2000
1900
1800
1700
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
Country #3
2000
1900
1800
1700
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
252
228
220
212
204
196
188
180
92
88
84
80
76
72
68
64
60
2100
2000
1900
1800
1700
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
38
Country #5
Country #4
Country #6
470
450
430
410
390
370
350
310
400
394
388
382
376
370
364
358
352
346
340
2300
2200
2100
2000
1900
1800
1700
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
330
2200
2100
2000
1900
1800
1700
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
Country #7
2300
2200
2100
2000
1900
1800
1700
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
Country #8
600
565
530
64
61
58
55
52
495
460
425
390
355
2000
1900
1800
1700
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
320
85
80
75
70
65
60
55
49
Country #9
2500
2400
2300
2200
2100
2000
1900
1800
1700
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
50
46
43
40
171
164
157
150
143
136
129
122
115
2100
2000
1900
1800
1700
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
39
Country #10
Country #11
2400
2300
2200
2100
2000
1900
1800
1700
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
1410
1340
1270
1200
1130
1060
990
920
850
6400
6100
5800
5500
5200
4900
4600
4300
2500
2400
2300
2200
2100
2000
1900
1800
1700
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
Expected Aggregate net demand for carbon emission quota units
MSEK/Mton
2000
1900
1800
1700
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
6000
6500
7000
7500
8000
8500
9000
9500
10000
10500
11000
11500
Emissions
Mton CO2
100
90
MAC-curve
MSEK/Mton
80
70
60
50
40
30
MAC(e^)
20
10
0
0
1
2
3
4
5
6
7
8
e^
Carbon dioxide emissions, Mton
9
10
12
e0
Fig. 1. Illustration of a country’s expected underlying net demand of carbon quota
units