A majority of the world’s electricity is generated using fossil-fired power plants, and many of these plants are aging coal power plants nearing obsolescence. The International Energy Agency projects energy consumption to increase globally by about 35 percent from 2010 to 2030, resulting in substantial investment in addition to replacing existing aging infrastructure. Simultaneously, governments and regional bodies are formulating carbon reduction targets.
For example, Europe has a near term goal of reducing its greenhouse gas (GHG) emissions by 20 percent from 1990 levels by 2020 and both the United States and Canada have set a similar goal of reducing GHG emissions by 17 percent from 2005 levels by 2020.To be achieved, these targets require a radical transformation in a wide range of industrial sectors, especially electricity generation, which is the largest single source of GHG emissions globally.
However, the nature and stringency of policies to reach carbon mitigation targets is still highly uncertain. Even if a chosen policy approach is established, policy uncertainty can still be significant. For example, carbon prices in Europe have varied from nearly 0 to 30 euros/ton of carbon dioxide since 2008 when the second phase of its cap-and-trade market started. In addition to carbon policy uncertainty, firms’ generation investment decisions are complicated because some carbon mitigation technologies, such as carbon capture and storage (CCS), are in the development phase and uncertainty exists regarding their cost-efficiency.
In this article, we synthesize the key results from three research projects, funded by Carbon Management Canada, that consider the impacts of carbon policy uncertainty in power generation investments at firm, market and economy wide levels.
Firm Level Insights
At the firm level, the impact of carbon policy uncertainty relates directly to the investment decision, i.e., on what power technology to invest in and at what time given that the investment needs to be made during the planning horizon. As is typical in practice, such investment models are driven by discounted cash flow based valuation.
We apply such an approach including (i) operational and accounting details in a yearly level until the end of the operating life time of the plant, (ii) carbon policy uncertainty via modeling plausible carbon prices using a probabilistic scenario tree, and (iii) firm’s risk aversion by constraining the net asset value at the 5th percentile of scenario outcomes, similar to the frequently used Value-At-Risk measure. To capture interactions among risks of existing and new plants, we include existing portfolio of power plants.
The most interesting result of this study relates to how the existing portfolio of power plants impacts the new risk-averse investment resulting in path dependency in investment decision-making. This impact is illustrated in Figure 1, which compares investment probabilities into a new power plant without existing plants and with a diversified portfolio of plants.
Figure 1 indicates that it is optimal for firms considering entry in the power market to delay their entrance and when they enter to invest in gas fired power plants (95 percent of scenarios) or wind farms (5 percent of scenarios). However, for a firm with a diversified mix of existing plants it is optimal to invest immediately in a coal fired power plant to hedge its risk exposure to low carbon prices due to already owned gas and wind power plants.
These results suggest that carbon policy uncertainty can create a path dependency in power plant investments depending on the GHG risk exposure of existing plants and their interaction with the new plants. For policy makers, these results suggest that path dependency should be addressed in the models used for analyzing the impacts of different policies. For firms, these results suggest that hedging evolving carbon policy uncertainty via portfolio diversification can provide a competitive advantage.
Whilst the firm level analysis allows including substantial amount of operational and accounting details, there are benefits of assessing power plant investments on market level as it makes it possible to capture competitive interactions among the firms.
Market Level Insights
At the market level, we consider an oligopoly of power producers and use a game theory model to capture firms’ interactions. The focus is again on how carbon policy uncertainty impacts power generation investments. To assess the impacts of carbon policy uncertainty we consider generation investments under known “deterministic” and uncertain “stochastic” policies.
The uncertainty is modeled as two possible policy outcomes, namely strict and relaxed, with associated probabilities. We include heterogeneity of firms in their risk aversion, where risk is measured as the lower projected profit in the two possible policy outcomes.
This study shows that firms invest more in both fossil and renewable technologies under stochastic carbon policy than under deterministic. This occurs because firms can partially hedge the policy uncertainty by being better prepared for both outcomes. Moreover, we show that capacity investments increase with uncertainty in carbon policy due to the same reason.
One of the most surprising results of this study, however, is that uncertainty in carbon policy, which is often considered to be harmful for market participants, can actually be beneficial for (i) consumers measured in expected consumer welfare, (ii) generation firms measured in their expected profits, and (iii) society measured in lower carbon dioxide mitigation costs.
These results hold when generating firms are risk averse and electricity demand is somewhat elastic over several decades when the power plants are operated. For policy makers, these results suggest that keeping some flexibility in deciding on the stringency of the carbon policy until appropriate information emerges and letting the market thus bear some policy uncertainty may not be bad after all.
Whilst the market level analysis allows including competitive interactions among the firms and investigating how consumers and firms are impacted by the carbon policy uncertainty, it does not allow answering macro level questions dealing with carbon dioxide emissions on economy wide level, for example. For that purpose we employ a macroeconomic model.
Economy Wide Insights
At the macro level, we consider technological change, firm behavior, and market clearing supply and demand for energy and energy intense goods and services across all sectors. Particularly, we consider the role of government policy uncertainty on CCS investments, which is being considered as one of the possible ways to meet the ever-stringent emission targets whilst still employing fossil fuel power generation.
The decision to invest in emissions-reducing technologies such as CCS is influenced by a number of factors, including policy signals and the cost of the technology. In particular, uncertainty about future carbon policy has the potential to dissuade firms from making the large capital investments required for CCS; beyond a few incomplete pilot projects, firms have not implemented CCS in spite of years of policy announcements from government. With hindsight, it appears they have thus far been accurate in anticipating future policy stringency will not be high enough to warrant investment in CCS.
Our results, derived from the CIMS3 model using recent cost estimates for CCS applications in the energy sector, suggest that confidence in future policy is enough to impact investment decisions and encourage early adoption of CCS technologies.
Moreover, confidence in increasing policy strength is likely to encourage the adoption of more expensive CCS technologies (e.g. post-combustion CCS in the electricity sector) in periods where policy stringency is below capture costs. The flip side is that uncertainty discourages early investment in CCS and results in greater overall policy costs to achieve a given emissions target.
In other words, substantial investment in CCS is unlikely to occur if policy stringency is much below $100/t for carbon dioxide equivalent emissions, or if firms believe it is unlikely that policy stringency will reach such levels in the future. To encourage CCS, policy therefore needs to be of sufficient stringency and certainty. For high cost and high effectiveness abatement like CCS, creating a stable policy environment is crucial for influencing investment decisions in infrastructure with life spans measured in decades.
At present, confidence (or lack thereof) in future policy could well mean the difference between achieving or missing near-term targets. In fact, it is highly unlikely that the 2020 targets will be met in Canada unless strong policies are implemented immediately with indication that they will increase in stringency in the future. Additionally, to achieve 2020 targets with the support of CCS technologies, more than just a carbon price will be required; governments will need to facilitate and encourage an accelerated regulatory approval and construction environment for CCS.
Our research shows that carbon policy uncertainty has significant impact on power generation investments and that these impacts can be different depending on which level of investment decision making is being considered. At the firm level, carbon policy uncertainty creates path dependency in resources acquisition with the result that new investment decisions depend on the existing power generation assets and how they interact with the carbon policy risk.
At the market level, carbon policy uncertainty incentivizes excess capacity investment in both fossil and renewable technologies, which over long run, given electricity demand is to some degree elastic, can be beneficial for both consumers and generating firms. At the economy wide level, our results show that sufficient long run policy stringency and certainty is needed in carbon policy to meet near and long term emission reduction targets, with a carbon price of $100/t for carbon dioxide equivalent emission or more.
Authors: Janne Kettunen is with The George Washington University, Chris Bataille is with Simon Fraser University, Liang Chen is with Navius Research and Noel Melton is with the University of Calgary.