Abstract
This study analyses factors related to allowance-trading behaviour for the first ten years of the existence of the European Union Emissions Trading System (EU ETS). Our empirical analysis employs a dataset that combines information on trading activities for more than 6000 companies with company characteristics. Indicators of trading activity include the volume and the number of transactions as well as the usage of intermediaries and of derivatives markets. For 2005–2014 and for the individual trading periods, we find that trading behaviour is related to the size of a company, its net position (the difference between free allocations and verified emissions), its sector affiliation, productivity, and location. We also find evidence that trading-related transaction costs affect trading activity in the EU ETS in all trading periods. Our results further suggest that net buyers (companies whose verified emissions exceed free allocations in a given year) are more likely to participate in emissions trading and to trade at higher volumes than net sellers are. We explain this asymmetry in behaviour—which might lead to a violation of Coase’s independence property—by potential asymmetries in the actual or perceived opportunity costs of holding allowances between net sellers and net buyers.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
The market stability reserve regulates the supply and demand of allowances depending on the number of allowances in circulation. It is expected to reduce the supply of allowances by at least 2000 million EUAs by 2023 (e.g. Bocklet et al. 2019) and to affect the time profile of low-carbon investments (Perino and Willner 2019).
Based on interviews with managers in the ceramics industry, however, Venmans (2016) concludes that companies which are short tend to abate more.
For multi-national companies, the national ultimate owner is the subsidiary that owns all other subsidiaries in a given country. In contrast, the global ultimate owner is the global parent company that ultimately owns all subsidiaries (usually by controlling the national ultimate owners). We refer to a company as the legal organization that operates closest to the regulated entity, i.e. the installation.
Trades of international offsets like certified emission reductions (CERs) and emission reduction units (ERUs) are included in the dataset as long as they take place within the EUTL registry. If an account registered within the EUTL acquires or transfers an offset certificate from a party not registered in the EUTL, the counterparty is usually not known. If both parties are registered in the EUTL, the transaction is treated like a normal transfer. For the period between 2008 and 2012 we are, in principle, able to identify the type of transferred units. Since 2013, this information is, however, no longer observable as offsets are converted into EUAs when imported into the system (banked offsets had to be converted as well).
This procedure extends work in previous studies (e.g. Jaraitė-Kažukauskė and Kažukauskas 2015; Hintermann and Ludwig 2019; Zaklan 2013) by using company registration numbers in addition to addresses and account names to match these datasets. This was possible after reporting of company registration numbers became mandatory in 2012.
We do not correct the data for carousel VAT tax fraud, which was particularly relevant in the second trading period (e.g. Frunza et al. 2011). However, since we cannot clearly identify trades for tax fraud purposes in the data, we decided not to correct our data. Because our analysis includes accounts only of companies with at least one regulated installation and where information in the ORBIS database was available, most of the fraud-related transactions should be excluded. Fraud-related transactions were typically carried out via private accounts owned by shell companies which were not related to a company covered by the EU ETS.
EUA futures are traded at regulated marketplaces (e.g. at the European Energy Exchange, Intercontinental Exchange). Participants must be registered, e.g. at the London Clearing House. The clearing accounts show heightened activity during a few days in December each year when futures are typically delivered. For our analysis, the following days showed significantly higher trade volumes (at least three times as high as on normal days) and are therefore selected to calculate the use of futures: 21–23/12 2005, 18–22/12 2006, 17–19/12 2007, 15–19/12 2008, 14–18/12 2009, 20–23/12 2010, 20–23/12 2011, 17–21/12 2012, 17–20/12 23/12 2013, 16–19/12 22–23/12 2014.
Forwards are traded bilaterally and not necessarily cleared. Delivery usually takes place during the last business day in November or the first business day in December. For our analysis, these days are: 30/11 01/12 2005, 30/11 01/12 2006, 30/11 03/12 2007, 28/11 01/12 2008, 30/11 01/12 2009, 30/11 01/12 2010, 30/11 01/12 2011, 30/11 03/12 2012, 29/11 02/12 2013, 28/11 01/12 2014.
This definition follows the literature (e.g. Jaraitė-Kažukauskė and Kažukauskas, 2015; Zaklan 2013) but abstracts from the fact that a company's verified emissions may depend on trading (versus abatement) activities and does not include banked allowances. As explained in “Appendix Compiling of the data”, due to inconsistencies in the data, we did not include information from the EUTL pertaining to banked allowances. Accounting for banked allowances would increase the value of net position, because companies which are short (or long) in year t are likely to have been short (or long) in year t-1. Because borrowing, unlike banking, is not allowed across trading periods, accounting for banking would asymmetrically affect the net positions of net buyers and net sellers.
The carbon-leakage list includes a large number of products from various industry sectors. In our case, these sectors include: refineries, iron and steel, metals, aluminium, cement and lime, glass and ceramics, pulp and paper, and chemicals as well as food, textiles, and machinery production.
The panel is unbalanced because of exit and entry of companies and installations, and because countries joined the EU ETS at different points in time. In particular, the three non-EU members, Norway, Iceland, and Liechtenstein as well as Romania and Bulgaria joined the system in 2007 and Croatia in 2013.
The \({\overline{x} }_{i}\) are referred to as Mundlak terms. They pick up the “between variation” and may be interpreted as the long-run effects. In comparison, the time-varying variables pick up the “within variation” and may be interpreted as the short-run effects. Because we worry, that the effects of unobserved heterogeneity may be correlated with the explanatory variables, our presentation and interpretation of the results will focus on the time-varying effects.
In general, the EU ETS allows banking of allowances across trading periods and years, but not borrowing across periods. For the first trading period, neither banking nor borrowing was possible across periods. Because allowances for the year t + 1 are allocated before companies need to surrender allowances for emissions for year t, borrowing is de facto feasible also across years within the same trading period. Thus, the period pertaining to banking of allowances is longer than the period for borrowing.
All findings that are not shown to save space are available from the authors upon request.
TAs were introduced only in 2013 and allow, in contrast to PHAs, trading in real time, while PHAs are delayed by 26 h (Art. 39.3 Registry Regulation No 389/2013). The majority of PHAs and TAs are opened by non-liable companies such as financial intermediaries as well as liable companies that use them to manage compliance and trading activities (Betz and Schmidt 2015; Cludius and Betz 2020). Some PHAs and TAs are held by non-governmental organizations or private individuals.
A more complete description of the processing of the EUTL data is available in Abrell (2021). The compiled data set is available for download under https://euets.info/background.
Abbreviations
- AOA:
-
Aircraft operator account
- CER:
-
Certified emission reductions
- CITL:
-
Community Independent Transaction Log
- CRE:
-
Correlated random-effects estimator
- EEX:
-
European Energy Exchange
- ERU:
-
Emission reduction units
- EUA:
-
EU allowances
- EU ETS:
-
European Union Emissions Trading System
- EUTL:
-
European Union Transaction Log
- GHG:
-
Greenhouse gases
- OHA:
-
Operator holding account
- PHA:
-
Person holding account
- TA:
-
Trading account
References
Abrell J (2021) Database for the European Union Transaction Log. https://euets.info/static//download/Description_EUTL_database.pdf. Retrieved: 14 May 2021.
Baudry M, Faure A, Quemin S (2021) Emissions trading with transaction costs. J Environ Econ Manag 108:102468
Betz RA, Schmidt TS (2015) Transfer patterns in Phase I of the EU Emissions Trading System: a first reality check based on cluster analysis. Clim Policy 16(4):474–495
Bocklet J, Hintermayer M, Schmidt L, Wildgrube T (2019) The reformed EU ETS—Intertemporal emission trading with restricted banking. Energy Econ 84:104486
Borghesi S, Flori A (2018) EU ETS Facets in the Net: structure and evolution of the EU ETS network. Energy Econ 75:602–635
Burtraw D, McCormack K (2017) Consignment auctions of free emissions allowances. Energy Policy 107:337–344
Cason TN, Gangadharan L (1998) An Experimental Study of Electronic Bulletin Board Trading for Emission Permits. J Regul Econ 14:55–73
Cludius J (2020) Ex-post investigation of cost pass-through in the EU ETS - an analysis for six industry sectors. Energy Econ 91:104883
Cludius J, Betz R (2020) The role of banks in EU emissions trading. Energy J 41(2):275–299
Cludius J (2018) Winners and Losers of EU Emissions Trading: Insights from the EUTL Transfer Dataset. Economics of Energy & Environmental Policy 7(2)
Coase RH (1960) The problem of social cost. J Law Econ 3:1–44
Cragg JG (1971) Some statistical models for limited dependent variables with application to the demand for durable goods. Econometrica 39:829–844
Ellerman AD, Convery FJ, de Perthuis C (2010) Pricing carbon: the European Union emissions trading scheme. Cambridge University Press, Cambridge
Engel C, Moffatt PG (2014) dhreg, xtdhreg, and bootdhreg: commands to implement double-hurdle regression. Stata J 14(4):778–797
Frunza M-C, Guegan D, Lassoudiere A (2011) Missing trader fraud on the emissions market. Journal of Financial Crime 18(2):183–194
Guo J, Gu F, Liu Y, Liang X, Mo J, Fan Y (2020) Assessing the impact of ETS trading profit on emission abatements based on firm-level transactions. Nat Commun 11:2078
Hahn RW, Stavins RN (2011) The effect of allowance allocations on cap-and-trade system performance. J Law Econ 54(S4):267–294
Hintermann B, Ludwig M (2019) Home Country Bias in International Emissions Trading: Indirect Evidence on Transaction Costs in the EU ETS. Working Paper, University of Basel
Jaraitė-Kažukauskė J, Kažukauskas A (2015) Do transaction costs influence firm trading behaviour in the European emissions trading system? Environ Resource Econ 62(3):583–613
Kahneman D, Tversky A (1979) Prospect theory: an analysis of decision under risk. Econometrica 47:263–292
Kahneman D, Knetsch JL, Thaler RH (1990) Experimental tests of the endowment effect and the Coase Theorem. J Polit Econ 98(6):1325–1348
Karpf A, Mandel A, Battiston S (2018) Price and network dynamics in the European carbon market. J Econ Behav Organ 153:103–122
Martin R, Muûls M, Wagner UJ (2015) Trading Behavior in the EU ETS. In: Gronwald M, Hintermann B (eds) Emissions trading as a policy instrument: evaluation and prospects. MIT Press, Cambridge, pp 213–238
Montgomery WD (1972) Markets in licenses and efficient pollution control programs. J Econ Theory 5(3):395–418
Mundlak Y (1978) On the pooling of time series and cross section data. Econometrica 46:69–85
Naegele H (2018) Offset credits in the EU ETS: a quantile estimation of firm-level transaction costs. Environ Resource Econ 70(1):77–106
Perino G, Willner M (2019) Rushing the impatient: allowance reserves and the time profile of low-carbon investments. Environ Resource Econ 74(2):845–863
Stavins R (1995) Transaction costs and tradeable permits. J Environ Econ Manag 29(2):133–148
Venmans F (2016) The effect of allocation above emissions and price uncertainty on abatement investments under the EU ETS. J Clean Prod 126:595–606
Zaklan A (2013) Why do emitters trade carbon permits? Firm-level evidence from the European emission trading scheme. EUI Working Paper RSCAS No. 19
Zaklan A (2020) Coase and cap-and-trade: evidence on the independence property from the European electricity sector. DIW Discussion Paper No. 1850
Acknowledgements
This research was funded by Stiftung Energieforschung Baden-Württemberg (Förderkennzeichen A 328 18). The authors thank two anonymous reviewers for their insightful comments and Jonatan Pinkse (University of Manchester) for his contribution to an earlier version of this paper.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix 1: Construction of the Database
1.1 Compiling of the Data
The Union registry is an electronic database managed by the European Commission that records all allowance transactions carried out under the EU ETS, including the allocation and surrendering of allowances, but also all transactions taking place between market participants. The European Union Transaction Log (EUTL) monitors, records, and authorizes all transactions occurring in the Union registry. Via the EUTL, the European Commission publishes data on allowance transactions as well as details from the Union registry. This information is now available on a three-year delay. The data can be downloaded free of charge (https://ec.europa.eu/clima/ets/).
In the EUTL transactions take place between registered accounts. All liable installations covered by the EU ETS are required to open an Operator Holding Account (OHA) for stationary installations or aircraft operator account (AOA) in the Union registry. In addition to these mandatory accounts, Person Holding Accounts (PHAs) and Trading Accounts (TAs) can be opened voluntarily in the Union registry for trading purposes.Footnote 15 Finally, a number of administrative accounts exist that belong either to the EU or to individual countries and are used for, amongst other procedures, the issuance, allocation, auctioning, or deletion of allowances.
On the account level, the EUTL includes information indicating the name of the account, the registry in which the account is registered, the related company registration number, and the associated account holder and installation. For account holders we know the name and address of the main account representative. For installations we know the type of activity and the address, and compliance data, including annual allocations, verified emissions, and surrendered allowances.
The EUTL records transfers of allowances between two accounts, providing information about the accounts involved, transaction types, transaction dates, and the number of allowances. Trading in futures and forwards is recorded only at the expiration date when a derivative is delivered to a buyer. The EUTL does not reveal information on prices per EUA or total payments.
Until 2012, a decentralized system of national registries existed. These registries were aggregated and checked in the Community Independent Transaction Log (CITL). In 2012, information was migrated from the individual registries to a single EU-wide registry and the CITL was replaced by the EUTL. In this context, all installations received new OHAs, i.e. all banked allowances had to be transferred from the old accounts to the new accounts, requiring a high number of internal transfers (see “Matching of Former and Current OHA”).
In addition to EUTL data containing ETS-related information, we also use financial data on the liable companies from the ORBIS database operated by Bureau van Dijk. We use financial data on the number of employees, revenues, industry classification (NACE), the company registration number and the home country of a company. To match the EUTL and ORBIS datasets, we relied primarily on the company registration number, the account name, and addresses of account holders (see “Matching of EUTL Accounts with Companies in the ORBIS Database”).
For the empirical analyses of transactions we consider only transactions involving OHAs, PHAs, or TAs. Thus, we do not consider transactions involving authorities such as the primary allocation or the surrendering of allowances. Because these EUA transactions are regulatory requirements rather than outcomes of deliberate decisions, they are not relevant in the context of this research.
After we matched the company information obtained from the ORBIS to the transaction-level dataset, we set up a panel dataset at the level of individual companies for our period of analysis. Some trades were carried out by accounts that we were not able to link to an ORBIS company; these involved mainly PHAs and TAs. However, we did not exclude these trades completely from our analyses. For example, if two PHAs traded with each other but only the transferor had a link to ORBIS, that transfer would be included in the transfer volume for that company in our dataset. Because the buyer did not have a link to ORBIS, however, this transaction could not be included on the buyer side. We believe that omitting these transactions does not significantly affect the results of our analysis, because the relevant company-level transaction volumes, which are the subject of our analysis, are not affected by this adjustment. Eventually, data on allocations, verified emissions, and surrendered EUAs were aggregated at the company level.
In total, we have 40,320 accounts in the initial list, of which 6466 could not be matched with ORBIS. The others were then aggregated to 15,014 companies. For our multivariate analyses the dataset includes fewer companies because we only include EUTL activities (but not aircraft operators) where NACE codes were available. We also excluded all observations where both verified emissions and allocations were zero, thus ensuring that plants that have ceased operations but are still listed in the EUTL were excluded. This leaves us with 8767 companies. Of those, information on the number of employees, sales, and profits was available for 6964 companies. Finally, we eliminated all observations where allocation exceeded verified emissions by a factor of ten to limit the effects of errors in the EUTL on our results. This left us with 6611 companies. Because of collinearity in the data matrix the samples available for the econometric analyses are somewhat smaller and vary across analyses.
We also tried to track company banking of EUAs but found substantial inconsistencies in the data when adding up the banked allowances over time. For example, in the wake of a reorganization of the EUTL in 2012, in many cases banked allowances from the second trading period appeared to not have been adequately transferred to the third trading period.
1.2 Matching of Former and Current OHA
The reorganization of the EUTL in 2012 led to new account types such as aircraft operating holding accounts. Hence, each installation needed to be associated with a new OHA. However, the EUTL provides the current OHA related to an installation only, not the OHA that was in place before the regime switch. To infer the previous OHA we used the following procedureFootnote 16:
-
1.
Matching the account name to the installation name and accepting matches if they are unique.
-
2.
Matching the account address to the installation address and accepting matches if they are unique.
-
3.
Matching on allocation information: In this stage, we use installation-level information on the amount of allowances allocated and surrendered and search for the corresponding transaction with the same amount of allowances and an administrative account of the respective registry involved. Again, only unique matches are accepted. We start with allocations followed by surrendering transfers.
This procedure allowed us to match more than 99% (i.e. 12,894 of 13,001) of the current OHAs to their OHAs before the regime switch.
1.3 Matching of EUTL Accounts with Companies in the ORBIS Database
Since 2012 operators of accounts are obliged to report a VAT registration number within the EUTL. This can be either a national or European VAT number. Because the ORBIS database also uses these VAT numbers a direct matching of accounts between the two databases based on the VAT number should be possible. However, because of reporting errors and differences in the formatting a direct matching was not feasible. We therefore use fuzzy matching based on a VAT number, the name of the account associated with that number, and the address of the account contact. These variables are used in automatic ORBIS batch searching using the account data as criteria for the search in the ORBIS database. Batch searching returns a number of possible matches together with the matching score. We then select the final match by inspecting the quality of the matches of the single fields.
Appendix 2: Descriptive Statistics
Appendix 3: Estimated Average Marginal Effects for Participation Decision
See Table 6.
Appendix 4: Results by Trading Period
Appendix 5: Results for Total Transactions with Asymmetric Response for Net Sellers and Net Buyers
Appendix 6: Results for Net Buyers and Net Sellers
Rights and permissions
About this article
Cite this article
Abrell, J., Cludius, J., Lehmann, S. et al. Corporate Emissions-Trading Behaviour During the First Decade of the EU ETS. Environ Resource Econ 83, 47–83 (2022). https://doi.org/10.1007/s10640-021-00593-7
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10640-021-00593-7