Three essays on power prices and the impact of green technology
- Autores: Shanshan Yuan
- Directores de la Tesis: Juan Ignacio Peña Sánchez de Rivera (dir. tes.)
- Lectura: En la Universidad Carlos III de Madrid ( España ) en 2021
- Idioma: español
- Tribunal Calificador de la Tesis: Rosa Rodríguez López (presid.), Manuel Moreno Fuentes (secret.), H. Torró (voc.)
- Programa de doctorado: Programa de Doctorado en Empresa y Finanzas / Business and Finance por la Universidad Carlos III de Madrid
- Materias:
- Enlaces
- Tesis en acceso abierto en: e-Archivo
- Resumen
Working towards a greener future has never been so important for us. The emission of greenhouse gases has resulted in significant shifts in weather patterns. Thus, the sustainable development of human society requires urgently the substitution for conventional energy sources. The transformation starts in the power industry first. To promote the use of renewable energy, many countries have set renewable energy target that calls for certain proportion of electricity consumption comes from renewable energy sources. For example, European Union (E.U.) set 35% target in 2030, United States has 50% target by 2035, United Kingdom aims at 100% by 2050. With favorable policies and financial supports, we have seen rapidly growth in renewable energy deployment. According to the International Renewable Energy Agency’s report, around 72% of the global electrical power expansion in 2019 was due to the development of renewables, mainly wind and solar (Hosseini, 2020). While more renewable generation would benefit the society in many ways, we haven’t learnt enough about the consequences.
Electricity, unlike other commodity, is special in terms of storability. Without feasible and economic way of large-scale storage, electricity must be consumed once it is generated. This fact indicates high spot price volatility (Escribano, Peña, and Villaplana, 2011). Moreover, the power market is quite inefficient with several large firms exercising market power, which leads to higher price (Hortaçsu et al., 2019; Borenstein, Bushnell, and Wolak, 2002). And market imperfections and regulation inadequateness can cause the “missing money problem” which refers to the fact that investments are below the optimal level (Winkler et al., 2016). Above these, renewable generation aggravates the situation by disrupting the market balance (Jónsson, Pinson, and Madsen, 2010; Paraschiv, Erni, and Pletsch, 2014; Kyritsis, Andersson, and Serletis, 2017) and increasing the price volatility (Edenhofer et al., 2013; Rubin and Babcock, 2013; Ketterer, 2014) due to its low marginal cost and intermittency nature. Then, these can lead to further consequences like higher risk premium required by the investors or less investment in power industry or higher cost to stabilize the grid etc. (Winkler et al., 2016; Edenhofer et al., 2013). As renewables’ influence will become greater over time thanks to its increasing market share, understanding more about renewables’ impact on the power market is critical for the market participants and policymakers. In this thesis, we intend to discuss the influence of renewables on power prices from three different perspectives.
This thesis contains three chapters. We start by discussing the asymmetric competition between renewable producers and conventional producers and the consequent change of hedging behavior of affected market participants. Then, we try to analyze the day-ahead price connectedness among regional markets in E.U. and how the renewable generation alters the price connectedness of one market with other markets. Lastly, we investigate the impact of renewables on the variance of electricity returns under different support schemes focusing on the Spanish intraday power market case.
In Chapter 1, “Ex-ante day-ahead premium and optimal hedging in electricity markets with green and brown producers”, we develop the equilibrium model on electricity forward premium proposed by Bessembinder and Lemmon (2002) (B&L from now on) by introducing the asymmetrical competition between the renewable producers and the conventional producers. Specifically, we focus on the day-ahead premium to avoid the market disruption from the financial traders. B&L model predicts that the volatility of expected spot prices increases the hedge pressure of retailers therefore decreases the forward premium, while the asymmetry of expected spot prices’ distribution grows the hedge pressure of conventional producers thus increases the forward premium. Yet, our results show that the ex-ante day-ahead forward premium is negatively or positively related to the variance and the skewness of expected spot prices depending on the demand level. Particularly, when demand is low, renewable generation has larger impact on the market, conventional producers then hedge more if they expect the spot prices to be more volatile, and retailers hedge more if they forecast more asymmetric spot prices with longer right tail. While, when demand is high, renewable generation’s impact is limited, high volatility of expected spot prices exposes the retailers with higher revenue risk, so they increase purchase in the forward market; high skewness of expected spot prices leads to a sharp increase in the conventional producers’ production cost, so they increase sale in the forward market. In addition, we discover the kurtosis of expected spot prices and the production share of renewable generation both are negatively related with the day-ahead premium, and the change in renewable supply curves positively affect the premium.
Using the trading data at unit level in the Spanish day-ahead market and intraday market, we derive the model-based expected spot prices. Analysis shows the main features of electricity spot prices are well accounted for by the model. Then, we apply panel fixed effect method on hourly data to test model implications. The empirical results largely support the model hypotheses.
We contribute to the existing literature by reconciling the mix empirical evidence found on the implications from B&L model and introducing new implications on the electricity forward premium. Also, we address the trade-offs of having more renewable generation. On one hand, renewable generation decreases the day-ahead premium. On the other hand, the variation of renewable supply will lead to more biased forward price. Thus, we believe the implications of our model are useful for practitioners and policymakers alike.
In Chapter 2, “The connectedness of European electricity day-ahead prices and the impact of renewable generation”, we focus on the impact of renewable generation on the price connectedness of 12 European power markets. The European Commission launched The Energy Union Strategy in 2015 targeting to build a clean, secure, and fully integrated European energy market. In the process of market integration, we shall expect that price shocks of one market can spillover to another market in the network. However, from Chapter 1, we see the renewable generation may interrupt the local energy supply and therefore influence the local power price. Thus, we ask the question of whether the renewable generation of one market will change its price connectedness with another market in the network.
To measure the price connectedness among selected E.U. power markets, we adopt Diebold and Yilmaz connectedness index (D&Y from now on) proposed by Diebold and Yilmaz (2009, 2012, 2014) based on the forecast error decomposition from the vector autoregression (VAR) model. If the markets are integrating and prices are converging, we shall expect an increasing trend in D&Y total price connectedness index over time. Also, D&Y measure can illustrate the directional connectedness. The “From” connectedness measures one market’s percentage of forecast error variance that can be explained by price shocks from other markets, while the “To” connectedness evaluates the total percentage of other markets’ forecast error variances that can be explained by one market. Thus, if we expect mutually spillover of price shocks among markets because of market integration, the price change of one market should increase both its “From” and “To” connectedness. Estimating the D&Y measure using t-lasso approach to account for the fat tails in the errors, we obtain a total connectedness of the day-ahead prices from 12 E.U. power markets at the value of 20.09%. Comparing to an average value of 45% in the studies on electricity price or volatility connectedness, our value suggests high trade barriers or large dissimilarities among the 12 power markets in the European Union.
After obtaining the dynamic “From” and “To” connectedness measures, we apply panel model on the data of the 12 power markets. We discover that the renewable generation decreases significantly both the “From” connectedness and the “To” connectedness of one market. This suggests that higher renewable generation would lead to less price connectedness of one market with other markets, which is consistent with the literature arguing that larger renewable generation can decrease price convergence (Woo et al., 2011; Horst Keppler, Phan, and Le Pan, 2016; Ardian, Concettini, and Creti, 2018). Meanwhile, we find no significant evidence suggesting that generation structure or cross-border transmission would affect the price connectedness of one market (Apergis, Fontini, and Inchauspe, 2017; Nepal and Foster, 2016), indicating other potential factors may explain the high pairwise directional price connectedness between markets identified in the chapter.
This study links the renewable generation with the power market integration by looking at the price connectedness among markets. We believe that this chapter is helpful for all market participants in understanding the spillover of price shocks and the impact of renewable generation on the market integration.
In Chapter 3, “The impact of renewables on the variance of power returns under different support schemes: the case of Spanish intraday power market”, we ask the question that whether the relationship between the renewable supply and the variance of electricity returns would be different under different renewable support schemes. According to our argument in chapter 1, the intraday position of renewable producers is affected by two factors: the prices, which indicate the arbitrage opportunities, and the supply conditions in general. Both factors for renewable producers can be influenced by the renewable support scheme (Ito and Reguant, 2016; Jägemann, 2014; Fagiani, Barquín, and Hakvoort, 2013). Winkler et al. (2016) argue that the renewable producers are less likely to accept extreme low price under market-oriented support scheme, leading to less renewables’ impact on the variance of electricity prices. Focusing on the Spanish intraday market, we intend to provide empirical evidence on this argument.
Using the trading data of the Spanish intraday market, we obtain the intraday renewable supply and intraday price from 2001 to 2017 covering the five periods of different renewable support schemes in Spain. The five periods are: the period from 2001 to 2004 (the initial Feed-in Premium scheme), the period from 2004 to 2007 (the combination of Feed-in Tariffs and Feed-in Premium scheme), the period from 2007 to 2012 (Feed-in Premium with cap and floor and cuts of tariff system), the period from 2012 to 2014 (the abolishment of renewable support), and the period from 2014 to 2017 (the new rate-based remuneration scheme). Thus, the Spanish renewable support scheme is becoming more market-oriented over time. For the method, we decide to apply the popular models for variance study in the literature, including both conditional variance model, GARCH and EGARCH, and unconditional realized variance measure with LS regression. Considering the fat tails in the series, we assume t-distribution for the errors and obtain more robust estimations. From the preliminary results, we find that the renewable intraday supply ratio, which is calculated as the renewable intraday supply over total renewable supply, is significantly and positively related with the variance of power returns. After introducing the support scheme dummies in the variance equation, we observe consistent and significant evidence showing that the renewable intraday supply increases the variance of power returns even more in the period of the abolishment of Feed-in Tariffs and Feed-in Premium support scheme during 2012 to 2014. This result points out the importance to minimize renewable policy uncertainty for stabilizing the power returns. Also, this indicates that renewable support schemes can decrease the impact of renewable supply on the variance of power returns, which is consistent with the literature (Ciarreta, Pizarro-Irizar, Zarraga, 2020). However, we observe no significant difference of the renewables’ impact on the variance of power returns comparing the Feed-in Premium support scheme, the combination of Feed-in Tariffs and Feed-in Premium support scheme, and the new rate-based incentive scheme in the Spanish power market.
In this chapter, we discuss the effect of the renewable supply on the variance of power returns under difference support schemes focusing on the case of Spanish power market, extending the literature on this topic (Winkler et al., 2016). For further study, looking at the unit level bidding behavior of renewable producers in detail may provide more insights on the question.
Centering on the question of how renewable generation will impact the power prices, we discuss and provide empirical evidence on the question from three different perspectives. We believe that understanding the influence of renewables comprehensively is essential for achieving a greener future and hope this thesis may deliver valuable insights on this topic.
