Previous studies estimating factors driving electricity demands in Indonesia overlooked cointegration issue by simply using OLS that may result in a spurious regression. This study uses ARDL method to estimate the influences of electricity price, oil price, income, along with urbanisation, number of electricity customers, and Asian economic crisis to electricity demands of different end-user sectors from 1969 to 2015. The estimation derives cointegrating long-run relationships in all sectors and the signs of error correction terms in the short-run models also meet the expectation. This study found that urbanisation as the highest contributor to rising electricity consumption in all sectors. Therefore, the government should accelerate the implementation of energy performance standards for home and office appliances. On the other hand, the government cannot effectively use price instruments to control consumption.
Electricity in all sectors cannot be substituted by oil fuel and, therefore, electricity consumption
is inelastic to its price. The electricity price is highly subsidised, especially for the residential sector and, therefore, the price should be adjusted gradually to avoid excessive electricity consumption.
Nevertheless, analysis in this chapter can be improved in several ways. First, future studies should use retail price for 34 electricity customer groups in all sectors, e.g., residential customers with subscripted capacity for 450 VA, 900 VA and 1300 VA. The use of average electricity revenue, which consists of the marginal price, block tariff, and fixed consumption charges, causes bias (Al Irsyad and Nepal, 2016, Alberini and Filippini, 2011). Second, using a measure other than HSD oil price could be a better substitute fuel price since it was not estimated in industrial and commercial demands due to the redundant variable problem. Third, an estimation by the fully modified OLS and bootstrap ARDL can provide more robust results in the presence of structural change, endogenous regressors, and inconclusive cointegration results (Adom and Bekoe, 2013, McNown et al., 2018).
RENEWABLE ENERGY PROJECTIONS FOR CLIMATE CHANGE MITIGATION: AN ANALYSIS OF
UNCERTAINTY AND ERRORS
(Published in the Renewable Energy)
4.1 Introduction
Increasing energy demand and the ongoing global push towards decarbonisation to mitigate the adverse effects of climate change have led to an intensification of renewable energy to stabilise emissions growth in the energy sector (Luderer et al., 2014, Dulal et al., 2013). Renewable energy has specific challenges such as higher investment costs, less reliable technology and intermittent supply issues. In light of these factors, political willingness profoundly influences the commitment to implement renewable energy targets by providing incentives, accepting higher electricity costs, settling contradicting policies and having weather-depended energy systems (Yi and Feiock, 2014, Jacobsson and Lauber, 2006).
Nevertheless, government commitment does not necessarily guarantee the achievement of renewable energy targets3. Intermittency and unreliable technology can cause overestimations as to the capacity factor (CF) of renewable energy. Such technical issues can become the main barriers to implementing proposals for 100% renewable energy supply (Delucchi and Jacobson, 2011, Jacobson and Delucchi, 2011, Heard et al., 2017, Lucas, 2017). For example, even China, the leader on renewable energy capacity, cannot maximise renewables-based electricity production as a consequence of grid connectivity problems and low-efficiency technologies (REN21, 2017, Wang et al., 2010).
One of the most common reasons for inaccuracies in energy projection is the use of incorrect assumptions (O’Neill and Desai, 2005, Linderoth, 2002). O’Neill and Desai (2005) and Winebrake and Sakva (2006) suggested that incorrect macroeconomic assumptions are the source of fossil energy projection errors. On the other hand, Gilbert and Sovacool (2016) viewed inappropriate policy analyses and wrong assumptions on capital costs and CF as the sources of renewable energy projection errors. Policy—influenced by economic, environmental and political factors, which vary in each country—along with other institutional
3 The setting of renewable energy targets is usually based on renewable energy projections; thus, the terms target and projection are used interchangeably in this thesis.
issues, determines the achievement of renewable capacity targets (Marques et al., 2011, Lauber and Jacobsson, 2016). Conversely, technical issues (e.g., the reliability of technology, efficiency, the intermittency of resources and CF) affect the achievement of electricity production targets.
Several studies have already analysed the accuracy of renewable energy projections, but the scope of their analyses is relatively limited for drawing broad conclusions. Gilbert and Sovacool (2016) focused on projections in the United States (US) Annual Energy Outlook (AEO), and thus their results could not capture global trends. In contrast, Metayer et al. (2015) analysed global-level projections in the World Energy Outlook (WEO) but ignored the effectual nullification caused by equality of failures and successes in projection implementation across each country. This chapter extends the scope of these previous studies to the US and to 27 European Union (EU) countries, each of which has a strong motivation for green electricity supply. The US is the country with the second largest total capacity for renewable energy, and the EU countries are recognised to have the highest per capita capacity of non-hydro renewable energy (REN21, 2017). In addition, dominant error sources are examined by comparing errors in projections of capacity and production of renewable energy.
Research questions in this chapter are as follows: What is the most achievable renewable energy target? What is the projection error ranges for different types of renewables?
Which error is dominant? Analysis in this chapter uses three indicators: mean percentage error (MPE), mean absolute percentage error (MAPE) and the mean of the difference between absolute percentage error of capacity and absolute percentage error of production (MDAPE).
These terms are defined in sub Chapter 4.3.2 on methodology. The contribution of this chapter is threefold. First, it guides policymakers to understand the uncertainties and errors in their renewable targets. Secondly, the results, by providing information about the most achievable renewable targets, may assist risk-averse countries to secure their energy supplies. Lastly, issues that need more attentions in renewable energy planning are identified. The remainder of Chapter 4 is organised as follows: sub Chapter 4.2 discusses previous studies on energy projection accuracy, sub Chapter 4.3 describes the data and methodology, and sub Chapter 4.4 presents the analysis results. Sub Chapter 4.5 discusses the implications of the findings for renewable energy policy, and sub Chapter 4.6 concludes the analysis.