Investing in finite-life carbon emissions reduction program under risk and idiosyncratic uncertainty

Highlights

• We use real option model to identify key features of CERP investment decision.

• We determine the optimal carbon price threshold to undertake a CERP.

• Investment decision is a non-monotonic function of idiosyncratic uncertainty.

• Increasing uncertainty until a moderate level can accelerate investment decision.

• Decreasing idiosyncratic risk can accelerate investment decision.

Abstract

This paper aims at emphasizing the ability of new frameworks of real option model to highlight key characteristics of industrial Carbon Emissions Reduction Program investment decision. We develop both theoretical arguments and numerical simulations with structural parameters calibrated on real-life data. We find that both radical uncertainty and risk lead to speed-up green investments, compared to the predictions of real option models that are normally used in green investment literature. The conventional “wait and see” attitude, questioned in recent developments of the real option theory, is not validated. In conclusion, our results should foster companies to implement green investments and help governments to define appropriate incentives to encourage green investments. Of particular note, the paper highlights that finance theory is not necessarily an obstacle to green investment decisions.

Introduction

While firms nowadays announce environmental concerns in their long-run strategy, evidences from the field show only few decisions for actively managing and reducing carbon emissions. Has very first willingness-to-pay been frozen at an early stage? Managers may have experienced much more expenses or less valuable profits than primarily thought. Operational and organizational costs for managing carbon emissions may be for instance only partially balanced by strategic competitive advantages vis-à-vis, say, green customers or socially responsible investors. Finally, uncertainty about Carbon Emissions Reduction Program (CERP) profitability seems to prevail in managers' behavior.

To explain disappointing results and the propensity of firms to postpone CERP, it is quite common to refer to standard implications of the classical real option theory (see Dixit and Pindyck, 1994, for a presentation): increased uncertainty on investment raises the value of waiting and thus decelerates investment. Recent advances in the real option literature have nevertheless questioned and renewed the standard approach (see the next section for an in-depth discussion). Summing up, it now appears that real option models may lead to far more ambiguous predictions on the relationship between uncertainty and the value of the option to wait than previously thought, when some traditional hypotheses are clarified, refined and/or extended. We therefore feel there is a need to clarify how such developments can impact traditional views about CERP, understood as CO₂-price related investment projects.

This paper uses recent accounts in the real option literature to explore whether and to which extent three dimensions related to characteristics of CERP can impact decision and timing of investment. The new literature suggests that these dimensions could individually alter the expected time to invest in CERP. These three dimensions are (a) the Knightian uncertainty, (b) the correlation between carbon prices and the market portfolio, and (c) the finite life of the project. By so doing, this paper also contributes to the real option literature by putting together, i.e. in a single framework, important issues that were only considered separately.

It turns out that the existence of a strong environmental uncertainty is widely recognized in the literature in green management (Busch and Hoffman, 2009). Uncertainty concerns the availability of resources in the future and ecosystem dynamics (Chichilnisky and Heal, 1998, King, 1995) and the long-term effect of polluting emissions on the ecosystem. It is also linked to changing social expectations, regulations and industry standards (Ishii and Yan, 2004, McDevitt et al., 2007, Husted, 2005). Risks arise from many sources affecting the cash flows of a project, price risks, technological risks and financial risks.

We will concentrate on only one source of risk and uncertainty namely the uncertainty about the CO₂ price. The relative youthfulness of the CO₂ management suggests there exists some degree of ambiguity (or equivalently Knightian uncertainty since Knight (1921)) about future prospects. Uncertainty about the future legal and regulatory framework may also be a source of concerns. Following the Kyoto protocol, European industrial businesses, about 12,000 plants, have to reduce their carbon emissions from the 1997 level of 20% in 2020. To achieve this objective, carbon allowances quantities are allocated to them and a European carbon market has been developed for the exchange of surplus or necessities. The decision of investing in CERP is clearly identified as linked to the carbon price. This market is currently lost in an over-abundance of allowances. A plan to rescue the carbon market will enter into force in 2021. The announce reform consists of introducing an automatic mechanism to control the market in order to prevent a price collapse. This carbon market was initially created in order to give a price that encourages industrials to invest in CERP.

Consequently, firms, managers and decision makers should hardly trust a given probability measure, such as the one the classical real option approach routinely considers, to assign a precise chance on future events they feel ambiguous. Some firms may even be quite averse to ambiguity, so that a pertinent decision-making model should account for ambiguity aversion. In this paper we address such concerns in a continuous time setting as follows. To deal with ambiguity, we consider that the firm has a set of probability measures over uncertain states in place of the single one traditionally assumed in the real option theory and, in order to model Knightian uncertainty, we use the k-ignorance specification, just like Chen and Epstein (2002) and Nishimura and Ozaki (2007).¹ To capture firms' ambiguity aversion, we employ the multiple prior approach that consists for managers to make decision on the basis of the worst case scenario. Other way saying, considered preferences are related to the maxmin expected utility of Gilboa and Schmeidler (1989).

The level of correlation between such markets and other capital markets is for instance only sparsely documented. Chevallier (2009) investigates the empirical relationship between carbon futures and macroeconomic risk factors. The author mainly demonstrates the statistical significance of the stock and bond market variables in explaining the variation of carbon futures prices. So, it is not obvious whether firms are aware of the kind of correlation and redundancy CO₂ markets can have with respect to capital markets. This important issue deserves to be clarified empirically but also theoretically. From a theoretical viewpoint, how large the idiosyncratic component of CO₂ price dynamics should be remains an open question. Dependence between financial asset prices and carbon prices exists because equity markets strongly focus on revenue effects of CO₂ prices (Bushnell et al., 2013). Investors consider the needs of carbon emission allowances of each firm and the revenue effect of carbon price moves on firm's revenue. Introduction of emission modifies operating costs of those industrial sectors covered by programs such as the Kyoto protocol or the European Emission Trading Scheme (Paolella and Taschini, 2008). These modifications have in turn a direct implication on share price. Overall, there is a strong link between the share prices of firms from the dirtiest industries (larger emissions) and CO₂ prices and a less important one for cleanest industries. Correlation between carbon markets and capital markets can also be analyzed through carbon prices drivers. According to Simshauser et al. (2007), main drivers of CO₂ pricing are economic growth for the demand side, cost of fuel and technology switching for the supply side. From a policy perspective, the extent of country participation and the aggressiveness of emission targets constitute key drivers. As these drivers are differently correlated with capital market, the correlation between carbon prices and capital market is complex and dependent upon the trading period.

If CO₂ dynamics is partly correlated with the capital market portfolio value, then ambiguity should be confined to the CO₂ idiosyncratic risk component. This is a key point to consider when designing Knightian uncertainty.

Finite life of CERP is the last characteristic we want to deal with. This issue that refers to fast-changing CERP technologies and associated relative obsolescence may have important consequences for decision makers. We expect this attribute to speed up the investment decision. It is indeed known that the value of the project decreases when its lifetime decreases (see Gryglewicz et al., 2008) in a setting with no ambiguity.

The paper is organized as follows. Section 1 makes a short literature review to discuss how the recent real option model can renew the way CERPs is understood. Section 2 delineates the model and characterizes the optimal investment trigger and the option value, examines how the optimal investment trigger responds to variation of the volatility, the ambiguity and the lifetime of the project. Section 3 presents the data used to calibrate the model. Section 4 offers numerical analyses to gain additional insight in the specific case of a carbon emission reduction investment. Section 5 concludes.