Do Carbon Permits Constitute a Respectable Asset Class? Efficiency, Liquidity and Volatility in comparison with Other Financial Markets

(1)

1 Do Carbon Permits

Constitute a Respectable

Asset Class?

Efficiency, Liquidity and Volatility in

comparison with Other Financial Markets

n

Morgan Hervé-Mignucci (

morgan_hm@yahoo.fr

), PhD student.

n

Jan H. Keppler (

jan.keppler@dauphine.fr

), Professor of Economics,

Université Paris-Dauphine.

Centre de Géopolitique de l’Energie et des Matières Premières, Université

Paris-Dauphine

(2)

2 Overview of emissions markets: CO

₂

n

Scope: approx. 12,000 CO

₂

-emitting installations across the EU split among utilities and

industrials

n

Asset traded: 1 EUA = the right to emit 1 ton of CO

₂

-equivalent

n

Two trading phase: 1st period (2005-2007) and 2nd period (2008-2012)

n

Current price: EUA 2007 = EUR 0.29 and EUA 2008 = EUR 24.00.

0

5

10

15

20

25

30

35 04/22/05

05/31/05

07/05/05

08/09/05

09/14/05

10/19/05

11/23/05

12/29/05

02/03/06

03/10/06

04/18/06

05/23/06

06/28/06

08/02/06

09/06/06

10/11/06

11/15/06

12/20/06

01/26/07

03/02/07

04/09/07

Vintage 2007 Vintage 2008

n

Aim: help GHG-emitting

installations curb their

emissions in a cost-effective

way through a

cap-and-trade mechanism

n

Regulatory framework:

Kyoto Protocol and EU

directive

Source: ECX

n

Three characteristics of interest: (1) Market crash in May 2006, (2) Phase I price

convergence towards zero and (3) phase I and II decoupling.

(3)

3 Series 1: CO

2 Allowance Spot - Powernext

PNX_P

PNX_LR

Mean

16.16248 -0.007543

Median

16.30000 0.000000

Maximum

29.75000 0.297252

Minimum

0.500000 -0.436816

Std. Dev.

8.577132 0.056345

Skewness

-0.449431 -1.351661

Kurtosis

2.021691 14.86608

Jarque-Bera

35.00666 2937.549

Probability

0.000000 0.000000

Observations

476

0 5 10 15 20 25 30 24/06/05 11/11/05 31/03/06 18/08/06 5/01/07 PNX P -0.6 -0.4 -0.2 0.0 0.2 0.4 24/06/05 11/11/05 31/03/06 18/08/06 5/01/07 PNX LR 0 10 20 30 40 50 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

Data sample: 24/06/2005 to 23/04/2007 - Source: Powernext

0 50 100 150 200 -0.375 -0.250 -0.125 0.000 0.125 0.250

(4)

4 Series 2: CO

2 Allowance 2008 - ECX

Data sample: 24/06/2005 to 23/04/2007 - Source: ECX

10 15 20 25 30 35 24/06/05 11/11/05 31/03/06 18/08/06 5/01/07 ECX08_P -0.3 -0.2 -0.1 0.0 0.1 0.2 24/06/05 11/11/05 31/03/06 18/08/06 5/01/07 ECX 08 LR 0 10 20 30 40 50 60 12 14 16 18 20 22 24 26 28 30 32 0 50 100 150 200 250 -0.3 -0.2 -0.1 0.0 0.1 0.2

ECX08_P

ECX08_LR

Mean

20.05080

-0.000559

Median

19.45000

0.000000

Maximum

32.25000

0.186526

Minimum

12.25000

-0.288246

Std. Dev.

3.964770

0.033691

Skewness

0.697527

-1.258423

Kurtosis

3.088401

18.29443

Jarque-Bera

38.83559

4775.053

Probability

0.000000

Observations

477

(5)

5 Comparison of asset returns over 1 year

Data sample: 21/04/2006 to 23/04/2007 – Sources: Powernext, ECX, CCX & Reuters

Dividends are excluded from asset return calculations for equity and equity indices.

-120%

-100%

-80%

-60%

-40%

-20%

0%

20%

40%

60%

CO2 - PNX Spot SO2 (EUR) - CCX CO2 - ECX 08 CO2 (EUR) - CCX Brent (EUR) POWER BASE CAL08 POWER PEAK CAL08 EUROSTOXX Oil Gas Sector Wilshire 5000 (EUR) EURUSD EUROSTOXX 50 CAC Small Cap 90 SBF 250 EUROSTOXX Utilities Sector Rhodia POWER BASE - Powernext EDF POWER PEAK - Powernext

(6)

6 Comparison of asset returns from 2005 to 2007

Data sample: 22/04/2005 to 23/04/2007 2007 returns are calculated yeartodate CO2 and EDF 2005 returns are calculated from inception

-Dividends are excluded from asset return calculations for equity and equity indices - Sources: Powernext, ECX, CCX & Reuters.

-200%

-150%

-100%

-50%

0%

50%

100%

150%

CO2 - ECX 08

CO2 - PNX Spot

SO2 (EUR) - CCX

CO2 (EUR) - CCX

SBF 250

EUROSTOXX 50

EUROSTOXX Utilities Sector

CAC Small Cap 90

Wilshire 5000 (EUR)

Rhodia

EDF

EURUSD

Brent (EUR)

POWER BASE CAL08

POWER PEAK CAL08

(7)

7

Comparative study

n

Sophistication degree of carbon exchanges

n

Volatility

n

Informational efficiency

n

Liquidity

n

Transaction costs

n

Arbitrage opportunities

n

Asset pricing

(8)

8 1. Sophistication degree of carbon exchanges

2005

2006

2007 YTD

SPOT

FUTURES

OPTIONS

02/05

Nord Pool

creation

10/05

Nord Pool

creation

06/05

Powernext

creation

10/06

ECX

creation

03/05

EEX

creation

10/05

EEX

creation

04/05

ECX

creation

01/06

Nord Pool

daily short

trading

allowed

05/06

Powernext

real-time

trading

Innovations to come:

§

New products: CER (announced for this year), ERU and AAU.

§

Standardized cross-asset products: clean spark and dark spreads.

12/06

ECX

min. tick

size

reduced

10/06

Powernext

launch

payment-on-delivery for

OTC contracts

(9)

9 2. Annualized volatility over 1 year

Data sample: 21/04/2006 to 23/04/2007 - common annualizing factor chosen = 260

Sources: Powernext, ECX, CCX & Reuters.

0% 50% 100% 150% 200% 250% EURUSD

CAC Small Cap 90

Wilshire 5000 (EUR)

SBF 250

EUROSTOXX Utilities Sector

EUROSTOXX 50 POWER BASE

-CAL08 (PNX)

POWER PEAK - CAL08 (PNX)

EDF

Brent (EUR)

Rhodia

SO2 (EUR) - CCX CO2 (EUR) - CCX

CO2 - ECX 08

CO2 - PNX Spot POWER BASE

-Powernext

POWER PEAK

(10)

10 2. Annualized volatility over 20 observations

Data sample: 24/04/2005 to 23/04/2007 - common annualizing factor chosen = 260

Sources: Powernext, ECX, CCX & Reuters.

0%

50%

100%

150%

200%

250%

300%

350%

400%

1/06/05

1/08/05

1/10/05

1/12/05

1/02/06

1/04/06

1/06/06

1/08/06

1/10/06

1/12/06

1/02/07

1/04/07

PNX CO2 Spot ECX 08 CO2 ccx (EUR) SO2 (EUR) DJES 50 Brent (EUR)

(11)

11 2. Annualized volatility over 20 observations

Data sample: 24/04/2005 to 23/04/2007 - logarithmic scale - common annualizing factor chosen = 260

Sources: Powernext, ECX, CCX & Reuters.

0% 1% 10% 100% 1000% 1/06/05 1/07/05 1/08/05 1/09/05 1/10/05 1/11/05 1/12/05 1/01/06 1/02/06 1/03/06 1/04/06 1/05/06 1/06/06 1/07/06 1/08/06 1/09/06 1/10/06 1/11/06 1/12/06 1/01/07 1/02/07 1/03/07 1/04/07

PNX CO2 Spot ECX 08 PNX BASE PNX PEAK

(12)

12

Unit root tests for prices

AIC ADF w/ intercept ADF w/ intercept and trend ADF w/ none PP w/ intercept PP w/ intercept and trend PP w/ none

CO2 Spot (Powernext) 2 -0,715 -2,653 -1,334 -0,653 -2,681 -1,275

CO2 2005 (ECX - EUR) 2 -2,386 -2,912 -0,414 -2,376 -2,870 -0,412

CO2 2006 (ECX - EUR) 2 -1,022 -2,424 -1,091 -1,094 -2,504 -1,086

CO2 2007 (ECX - EUR) 2 -0,592 -2,467 -1,240 -0,635 -2,525 -1,235

CO2 2008 (ECX - EUR) 2 -2,572 -3,465 -0,756 -2,478 -3,347 -0,758

CO2 2009 (ECX - EUR) 2 -2,629 -3,405 -0,729 -2,517 -3,270 -0,729

CO2 2010 (ECX - EUR) 2 -1,240 -2,250 0,183 -1,732 -2,650 -0,040

CO2 (CCX - USD) 0 -1,516 -1,520 0,404 -1,528 -1,576 0,387

SO2 2007 (CCX - USD) 9 -1,108 -2,130 -0,733 -0,672 -1,866 -0,682

EDF (EUR) 2 -0,579 -2,154 1,934 -0,483 -2,093 2,035

DJ Eurostoxx 50 (EUR) 10 -0,544 -2,355 1,984 -0,720 -2,743 1,766

Wilshire 5000 (USD) 12 -0,087 -1,909 1,986 -0,407 -2,412 1,707

Base spot (Powernext) 8 -3,543 -3,915 -1,425 -5,763 -6,276 -2,028

Peak spot (Powernext) 9 -3,956 -4,281 -1,723 -6,515 -6,917 -2,682

Base CAL08 (Powernext) 2 -1,540 -1,027 0,462 -1,570 -1,140 0,396

Peak CAL08 (Powernext) 3 -1,659 -0,696 0,753 -1,642 -0,704 0,749

Brent Spot (IPE - EUR) 3 -2,210 -2,211 0,163 -2,329 -2,332 0,141

EURUSD 0 -0,669 -2,730 1,101 -0,530 -2,628 1,198

Signif. @ 90, 95 and 99% Signif. @ 90% and 95% Signif. @ 90%

Not significant @ least @ 90%

ADF - Exogenous factors PP - Exogenous factors

3. Informational efficiency:

unit root tests

§

Emissions prices follow a non-stationary process

(13)

13

Unit root tests for 1st difference

AIC ADF w/ intercept ADF w/ intercept and trend ADF w/ none PP w/ intercept PP w/ intercept and trend PP w/ none

CO2 Spot (Powernext) 2 -10,010 -10,010 -9,954 -16,415 -16,418 -16,391

CO2 2005 (ECX - EUR) 2 -6,000 -5,978 -6,021 -8,230 -8,200 -8,262

CO2 2006 (ECX - EUR) 2 -9,367 -9,388 -9,343 -14,787 -14,794 -14,782

CO2 2007 (ECX - EUR) 2 -10,424 -10,442 -10,379 -16,627 -16,632 -16,605

CO2 2008 (ECX - EUR) 2 -11,081 -11,070 -11,089 -18,635 -18,617 -18,651

CO2 2009 (ECX - EUR) 2 -11,018 -11,007 -11,027 -19,068 -19,049 -19,084

CO2 2010 (ECX - EUR) 2 -4,675 -4,768 -4,691 -8,281 -8,412 -8,336

CO2 (CCX - USD) 0 -23,519 -23,524 -23,497 -23,454 -23,459 -23,452

SO2 2007 (CCX - USD) 9 -6,150 -6,199 -6,144 -19,143 -19,171 -19,151

EDF (EUR) 2 -10,722 -10,711 -10,423 -17,697 -17,675 -17,506

DJ Eurostoxx 50 (EUR) 10 -6,752 -6,750 -6,450 -22,571 -22,552 -22,433

Wilshire 5000 (USD) 12 -6,796 -6,833 -6,460 -21,490 -21,482 -21,375

Base spot (Powernext) 8 -9,239 -9,228 -9,247 -28,138 -28,103 -28,170

Peak spot (Powernext) 9 -8,990 -8,978 -8,998 -27,462 -27,428 -27,495

Base CAL08 (Powernext) 2 -12,743 -12,810 -12,739 -23,638 -23,695 -23,644

Peak CAL08 (Powernext) 3 -12,180 -12,378 -12,128 -25,630 -25,830 -25,590

Brent Spot (IPE - EUR) 3 -10,918 -10,907 -10,921 -23,042 -23,019 -23,060

EURUSD 0 -22,150 -22,155 -22,113 -22,216 -22,226 -22,167

Signif. @ 90, 95 and 99% Signif. @ 90% and 95% Signif. @ 90%

Not significant @ least @ 90%

ADF - Exogenous factors PP - Exogenous factors

3. Informational efficiency:

unit root tests

§

Emissions prices 1

st

_{differences follow a stationary process: spot and Vintage}

CO

₂

prices are first-order integrated, I(1).

§

These results for EUA prices are as expected consistent with Albrecht, Verbeke

and De Clerq (2005) efficiency analysis for the US SO

₂

market: “The SO

₂

permit

market is basically as efficient as financial markets”.

(14)

14 3. Informational efficiency: phase I and II prices

0

5

10

15

20

25

30

35

04 /22/05

05 /31/05

07 /05/05

08 /09/05

09 /14/05

10 /19/05

11 /23/05

12 /29/05

02 /03/06

03 /10/06

04 /18/06

05 /23/06

06 /28/06

08 /02/06

09 /06/06

10 /11/06

11 /15/06

12 /20/06

01 /26/07

03 /02/07

04 /09/07

Vintage 2007

Vintage 2008

Source: ECX

UK entitled to

challenge NAP

I decision

Guidelines

for NAP II

published

First national

reports for 2005

emissions

2005 emissions

official release

NAP II review

for 11 MS

announced

9 NAP II cuts, only 1

accepted - interphase

banking forbidden

3 major NAP

II reviewed

2006 emissions

registry open but

(15)

15 4. Liquidity: monthly historical volumes

1

1 In Mt - Data from ECX, Powernext, Nord Pool, EEX, EXAA and Point Carbon OTC deals estimation - year-to-date data for 2007

Source: Point Carbon

262 Mt

817 Mt

540 Mt

+ 212%

(16)

16 4. Liquidity: volumes and turnover

Data source: Point Carbon and Euronext

n

EUA total volume estimate for 2007 is 1,550 Mt (Point Carbon). Forecasted

turnover for 2007 is 0.72x.

n

For comparison purpose, turnover for oil is approximately 4.00x in 2004.

CO

₂

(Mt)

EDF (nb shares)

Volume 2005 since inception

262 133 134 345

Volume 2006 FY

817 399 508 226

Volume 2007 YTD

540 122 470 390

Turnover 2005 since inception

0.12x

1.03x

Turnover 2006 FY

0.38x

2.89x

Turnover 2007 YTD

0.25x

0.94x

(17)

17 4. Liquidity:

bid-ask spread for spot and 2008 contracts

0,00 0,10 0,20 0,30 0,40 0,50 0,60 0,70 0,80 0,90 1,00 1,10 18/02/06 4/03/06 18/03/06 1/04/06 15/04/06 29/04/06 13/05/06 27/05/06 10/06/06 24/06/06 8/07/06 22/07/06 5/08/06 19/08/06 2/09/06 16/09/06 30/09/06 14/10/06 28/10/06 11/11/06 25/11/06 9/12/06 23/12/06 6/01/07 20/01/07 3/02/07 17/02/07 3/03/07 Bid-Ask Spot 0,00 0,10 0,20 0,30 0,40 0,50 0,60 0,70 0,80 18/02/06 4/03/06 18/03/06 1/04/06 15/04/06 29/04/06 13/05/06 27/05/06 10/06/06 24/06/06 8/07/06 22/07/06 5/08/06 19/08/06 2/09/06 16/09/06 30/09/06 14/10/06 28/10/06 11/11/06 25/11/06 9/12/06 23/12/06 6/01/07 20/01/07 3/02/07 17/02/07 3/03/07 Bid-Ask 2008 Source: Argus Source: Argus

(18)

18 5. Transaction costs

n

Country-level:

¨

transaction costs varies upon Member State

application of the Directive and organization (EEA

report, 2007).

n

Carbon exchange-level:

(19)

19

0

1

2

3

4

5

6

7 1/01/078/01/07

_15/01/07

_22/01/07

_{29/01/075/02/07}

_12/02/07

_19/02/07

_{26/02/075/03/07}

_12/03/07

_19/03/07

_{26/03/072/04/079/04/07}

_16/04/07

_23/04/07

_{30/04/077/05/07}

_14/05/07

_21/05/07

_{28/05/074/06/07}

CO2 spot

(Powernext)

5. Transaction costs

n

Insight from current spot and vintage 2007 prices

¨

As a result of a long market, spot and current vintage

price converge toward zero.

0 5 10 15 20 25 30 35 24/06/0524/07/0524/08/0524/09/0524/10/0524/11/0524/12/0524/01/0624/02/0624/03/0624/04/0624/05/0624/06/0624/07/0624/08/0624/09/0624/10/0624/11/0624/12/0624/01/0724/02/0724/03/0724/04/0724/05/07 CO2 spot (Powernext)

0.20 - 0.25

(20)

20 6. Arbitrage opportunities

Intra-market arbitrages over time

Cross-market arbitrages

n

Spot prices exhibited basis

convergence for Vintage 05 and 06.

n

High correlation between spot

and phase I vintages (2005-2007).

n

Futures prices analysis implies

standard no-arbitrage pricing

(Daskalakis, Psychoyios & Markellos,

2006).

n

Correlation between phase 2 EUA

contracts (2008-2012) and Calendar

power contract for 2008.

n

Arbitrages between CO

₂

prices

and theoretical CO

₂

switch price.

(21)

21 7. Carbon assets pricing: a pre-crash survey

n

Benz and Trück (2006) review the stylized facts of emission allowances and proposes modeling

spot prices behaviour with regime-switching models that enable to capture the characteristics of

CO2 allowances.

n

Daskalakis, Psychoyios and Markellos (2006) propose an empirical study of CO2 prices using

spot and future prices from the EEX. They found that spot prices can be well approximated by a

jump-diffusion model and that futures prices implies that the market participants adopt standard

no-arbitrage pricing.

n

Seifet, Uhrig-Homburg and Wagner (2006) propose a “tractable stochastic equilibrium model

reflecting stylized features of the EU ETS and analyze the resulting CO2 spot price dynamic”.

They found that an adequate CO2 process “does not necessarily have to follow any seasonal

pattern”, ”should possess the martingale property” and “exhibit a time- and price-dependent

volatility structure”.

n

Uhrig-Homburg and Wagner (2006) prepared a survey on CO2 price drivers and CO2 derivatives

expectations. They also present an expert survey on CO2 prices risk management, CO2

derivatives availability and design properties. Uhrig-Homburg and Wagner find there are chances

of success for futures and forward on CO2 and for options as well. They propose a

European-style standardized CO2 option with a cap set at EUR 40.

(22)

22 Summary table

* indicates a highly changing volatility

CO2 Spot CO2 2008 SO2 2007

CO2 (CCX) 2007 Wilshire 5000 Eurostoxx 50 EDF Power spot - base Power spot - peak Power fut - base Power fut -

peak Brent EURUSD

Sophistication degree Futures and options Futures Tracker, futures and options Tracker, futures and options Spot Spot, futures and options Spot, futures and options

Volatility (average over sample

period) High* (29%) High* (16%) High (16%) High* (21%) Low (6%) Low (6%) Low* (8%) Very high (138%) Very high (162%) Low (7%) Low (7%) Average (14%) Very low (3%)

Informational efficiency I(1) I(1) I(1) I(1) I(1) I(1) I(1) I(1) I(1) I(1) I(1) I(1) I(1)

Liquidity High High

Transaction costs Low Low

Arbitrage opportunities Physical arbitrages might exist Cross-rate Asset pricing Various models tested Various models tested

Spot, futures and

options Spot and futures

Carbon markets volume grow to more mature level already achieved among

financial and commodity markets

High except for some ETFs High

As markets develop, transaction costs are

being reduced: estimate for CO2 circa EUR 0.25

Low except for some ETFs Low

Cross-market arbitrage opportunities might

exist In case of mispricing

Arbitrages depends on interconnection

capacity and production ability of the

arbitrageur.

Regime-switching and jump-diffusion

(23)

23 Concluding remarks

n

Carbon permits constitute a proper asset class.

n

Quantitative models alone are not sufficient to

explain carbon prices movements because:

¨

The main drivers are often changing when trying to

explain historical price evolution and

¨

Highest price movements are led by regulatory

announcements.

n

Stabilization of the institutional framework is key

to carbon market future success.