This paper presents a model for assessing economic losses caused by electricity cuts as well as Willingness-to-Pay to avoid these power outages as an approximation to the value of supply security. The economic effects for simulated power cuts from 1 to 48 h, which take the affected provinces, the day of the week and the time of day into consideration, can be calculated using the assessment tool APOSTEL. The costs due to power cuts are computed separately for all sectors of the economy and for households. The average value of lost load for Austrian households and non-household consumers in the case of a power cut of 1 h on a summer workday morning was calculated to be 17.1 € per kWh of electricity not supplied.
In the past decades, Europe has experienced an unprecedented degree of electricity supply security.1 Nevertheless, the current status of reliability should not distract from the fact that the future development of electricity supply security is uncertain as production as well as distribution experiences significant restructuring. This transformation is taking place at three levels potentially affecting security of supply:
Firstly, challenges arise due to changes in the market framework or as a consequence of deregulation and unbundling imposed by EU directive 2003/54/EC (European Commission, 2003).2 Secondly, the significant growth of electricity generation from renewable energy sources implies increasing levels of supply volatility thereby putting pressure on transmission and distribution systems (Borggrefe and Nuessler, 2009 and Boxberger, 2005, or BDEW, 2011). Thirdly, the current and anticipated growth of electricity consumption in developed countries such as Austria3 requires capacity enhancements and innovative solutions. From a technical and public acceptance perspective these infrastructure measures become increasingly difficult to implement (Netzentwicklungsplan, 2012).
Together, these developments represent significant challenges to the power infrastructure and to the preservation of the current level of electricity supply security in the future.4
Selection and design of the appropriate measures for addressing these challenges require knowledge about their costs and their benefits. Hogan (2008) and Eto et al. (2001) discuss issues of the electricity market structure in the United States and find the necessity to assess the economic value of supply security enhancing measures as complement to the evaluation of their technological benefits. This is an essential prerequisite for regulatory policy and for the justification of investment decisions. While efficient decision making regarding security investments is hampered by the lack of precise knowledge of the benefits of potential enhancing measures, large scale failures of the power system are supposed to have increasingly serious consequences for the society. In the presence of near-perfect levels of supply reliability and increasing electricity dependence, societies are getting more vulnerable to power outages as preparation for prolonged outages becomes more difficult and less of a concern. This is known as the double paradox, researched in detail by Luiijf and Klaver (2000) and De Hoo et al. (1994) for the Netherlands.
Despite their increasing dependence on uninterrupted electricity supply, consumers send hardly any signals about their valuation of energy supply security to suppliers, who thus misinterpret the benefits of reliability improvements and postpone infrastructure investments (Böske et al., 2007). In the special case of grid-bound supply systems, such as electricity,5 customers have for physical reasons no option of choosing an operator with a more adequate level of supply security. In addition to these specific economic aspects of electricity supply security, the short- and medium-term resilience of infrastructures in spite of security-preserving investments not being made creates incentives to further postpone investments. Precise knowledge of the importance of uninterrupted electricity supply to society is thus paramount. This research aims at providing an economic assessment of the value of electricity supply security which can be used – among others – for energy political decisions, benefit cost analyses, or the design of regulatory frameworks.6
Since electricity supply security constitutes a non-market good, which can only be purchased in combination with the physical product (electricity), its value cannot be elicited by market transactions (Kariuki and Allan, 1996a). That is why usually the effects of a failure of electricity supply are utilized for the value elicitation of service reliability (Baarsma and Hop, 2009 and De Nooij et al., 2007, or Woo and Pupp, 1992, for instance). In this study the costs of power cuts to non-household consumers, which include businesses, public sector entities and non-government organizations (NGOs), along with the Willingness-to-Pay (WTP) of household consumers to avoid power outages are analyzed as a proxy of the value of security. With the economic assessment tool presented in this paper it is possible for the first time to collect data on the power outage costs for different consumer groups and to simulate the effects of power outages from one to 48 h in Austria.
This paper proceeds as follows: Section 2 describes the different outage cost categories and introduces the methodology utilized to assess the losses due to power cuts. Section 3 contains the results from the elicitation of non-household consumers' outage costs and households' WTP to avoid power interruptions. The assessment tool and its application for a power outage case study (simulated 12‐hour power cut in Austria) are presented. Section 4 summarizes and adds a conclusion concerning the need for further research.
Subsequent to the primary survey on the vulnerability of businesses, public sector entities and NGOs, the individual non-household consumer's losses in the case of a power outage on the basis of value-added model and the survey inputs had to be standardized in order to extrapolate these outage costs to the national level. Complete aggregates of sectors or regions were formed subsequently. To extrapolate in this way, the share of total losses as percentage of the average daily added value of a single non-household consumer was derived. The aggregate of all shares was then adapted to the entire economic sector proportionally to the electricity consumption of the respective economic branch as well as the time-specific and regional characteristics of the power cut under investigation. The amount of electricity not supplied (in GWh) for the sectors under consideration is reported. Non-household consumers identified to be severely or very severely affected by a power outage, and the number of persons employed there, are listed. The value of lost load (VoLL) is presented for every economic subsector and for households. The VoLL depends on the outage characteristics but lies in the range of 0.7 €/kWh (water supply; sewerage; waste management and remediation activities) to 39.9 €/kWh (construction).
For residential customers the methodology described in Section 2.2 and survey-based data were utilized to elicit households' WTP10 to avoid power outages via a discrete choice estimation model. A mean WTP of 17.3 € to avoid a 24-hour power outage was found. To avoid a 12-hour power cut, households are willing to pay 9.9 €. Households' WTP for the prevention of a 4-hour power outages amounts to 3.8 €. WTP to avoid a 1-hour power cut was assessed at 1.4 €. WTP independent of the simulated outage's duration is 33.4% higher in winter than in summer. Analogously to non-household consumers, the number of households severely or very severely affected and the number of persons living in these households are presented in the assessment tool. As with non-household consumers, the shortfall of electricity is derived from synthetic load profiles. The VoLL, the mean economic loss per hour of outage and the economic loss per hour of outage per member of household are listed. Unexpectedly, it seems to make no statistically significant difference whether advance warning of a power outage is given or not. While also the coefficient age is not statistically different from zero, the variables season, size of the outage area, participants' sex, education and household income are highly significant. The estimation results are provided in Appendix A2.
3.1. Macroeconomic outage costs assessment
Using the hitherto described assessment approach, a macroeconomic assessment model capable of simulating power outage-related economic losses of households and non-household consumers in conjunction was developed.11 The Austrian Power Outage Simulation Tool of Economic Losses (APOSTEL) allows for the analysis of the effects of hypothetical power outages for each of the nine Austrian provinces and for 15 economic sectors, as well as for households in a predefined region, at an arbitrary day. Seven indicators for the economic impacts of power outages on household and non-household consumers are calculated. Direct as well as indirect economic losses resulting from power outages are accounted for. However, APOSTEL only simulates the resulting blackout, independently of what caused it. Costs related to the damage to or resulting from the destruction of electricity infrastructure – which nevertheless occur in most power outages – are not assessed because this cost category highly depends on the outage cause and its attributes.
The central outage cost figure reported in many studies, the VoLL, significantly depends on the characteristics of the power cut under consideration. We thus carried out a sensitivity analysis of the VoLL. Table 2 shows that during a 1‐hour weekday power outage in Austria, the macroeconomic VoLL (on average household and non-household consumers) was calculated to be 3.2 € per kWh (summer evening) and 21.2 € per kWh of electricity unsupplied (winter morning), respectively. The VoLL shows a negative correlation with respect to the duration of the power cut, which implies declining marginal outage costs.3.2. Power outage case study
In this section a power outage and its effects in terms of economic damage, affected household and non-household consumers as well as the kWh of electricity not supplied are analyzed. Table 3 summarizes the effects of this 12‐hour power outage, which takes place on a summer workday (in this case August 16th 2011) and affects all Austrian provinces. The large non-household share of the total losses is due to the outage time, which takes place on a workday at 10 a.m.The sector-specific assessments of the outage costs and of the social effects in terms of people and non-household consumers affected due to an outage of this kind are presented in Table 4. Nearly all Austrian households and non-household consumers are affected.Table 5 shows three additional economic indicators of the damages of the power outage under consideration. As the losses of certain economic activities are partly due to stoppages in upstream sectors, there is a tendency to overestimate the VoLL for those downstream economic activities which are particularly dependent on the functioning of basic infrastructures.12 Thus, APOSTEL additionally reports the average loss per outage hour and the average loss per hour per employee. This latter figure does not directly depend on the energy intensity of the sector in question, and is thus well suited to compare one sector with another. If the absolute loss is related to the individual employee in the sector in question, a particularly high value is obtained if this sector employs relatively few people and incurs high economic losses.13 Thus, in most sectors this indicator counterbalances the VoLL.Non-household consumers' losses in the case of a power outage strongly depend on the characteristics of the outage (time of day, season, duration, etc.). In the case of the analyzed 12-hour power outage on a workday, non-household consumers' total losses amount to almost 450 million €. This corresponds to about 0.15% of the annual Austrian GDP. Due to the correlation of value adding processes and electricity consumption almost all productive activities are affected in the case of a widespread power outage. In Appendix A3 the central results from this study are compared to the findings of relevant international studies using the main indicator of electricity supply security's value (VoLL).14
