Análisis del impacto de las políticas climáticas europeas en la competitividad de la industria española

Resumen

This thesis evaluates the impact of climate policies established by the European Union on the competitiveness of the Spanish economy, focusing on the European Union Emissions Trading System (EU ETS). A key element of this evaluation is to assess the degree to which the opportunity cost of CO2 emissions is considered by the different industries. Indeed, although the usual assumption is that industries value their CO2 emissions at market prices, regardless of how they acquired the emission permits, there may be an "endowment effect", that is, that the opportunity cost may depend on the emissions being assigned by the allocation method. Transaction costs may also cause similar deviations. The analysis takes into account the effect of the allocation of emissions on business decisions. In order to achieve these objectives, an integrated model of the Spanish industry was constructed and validated to estimate the behavior of the industry under the European climate policy. This tool is a partial-equilibrium, bottom-up linear optimization model that integrates five of the most emission-intensive industrial sectors in the EU ETS: steel, cement, oil refining, tiles and bricks, and electricity generation. For each sector, the different production processes and technologies are defined as well as the alternatives currently available to improve processes and hence contribute to reducing GHG emissions and promoting energy savings. The model calculates the optimal strategy to meet demand at the lowest cost given different emission reduction scenarios compared to the baseline situation (Business As Usual, BAU). It determines the optimal combination of internal abatement possibilities of emissions from each sector, covering, in this way, a greater level of detail. Such integration was carried out in GAMS (General Algebraic Modeling System), a high-level modeling system for mathematical modeling problems. The results are presented as greenhouse gas marginal abatement costs curves (MACC) for major emitting industrial sectors. These curves allow to know the marginal cost to reach a given reduction target or the reduction expected for a certain price of CO2. The thesis also evaluated the behaviors of the agents by analyzing the actual transactions between installations participating in the EU ETS. If agents behave rationally, they should sell permits as long as their marginal abatement costs are lower than the market price, and buy permits if their marginal abatement costs are greater than the price of the permit (and if they have the capacity to increase production). By observing the actual transactions of the facilities and knowing their marginal abatement costs, it is possible to evaluate if there is any bias or deviation in this estimation of the opportunity cost. This thesis and the results presented allow to put forward two conclusions that are very relevant for the current regulation and future design of the European Union ETS but also for other carbon markets across the world. First, when assessing policies that affect carbon markets, it was found that the market should be modeled in an integrated way, accounting for all the relevant sectors and their interactions. As mentioned before, the model allows for not imposing separate (and somehow artificial) mitigation shares for each sector, but instead represents the real behavior of a market: least cost options across all sectors are identified, and the interaction between sectors taken into account. The second conclusion has to do with the behavior of the agents participating in the ETS market. The results show, in general terms, behaviors consistent with economic rationality for the power sector: agents sell allowances when it is cheaper to abate emissions internally or when they have over-allocation, and they buy them when the allowance price is lower than their abatement cost. According to the model and the assumptions made, the other sectors analized show non-rational behaviour in analizados muestran un comportamiento no racional en algunos aspectos de acuerdo al modelo implantado y a las estimaciones realizadas de los datos de entrada.