**4. Conclusions**

Australia has an international obligation to reduce its CO2 emissions under the Paris Agreement, and Australia's NEM is a CO2-intensive power system by OECD standards [52]. In this article, a requirement to lower the CO2 intensity of the NEM has been taken as an exogenous policy constraint in the form of a 40% VRE market share. The purpose of this modeled constraint was not to test optimal levels or the composition of VRE plant, nor the structural adjustment task and welfare implications of exiting coal plant at-scale. The specific line of inquiry was to determine whether the wide-spread use of *o*ff*-market* government-initiated CfDs to achieve higher VRE and lower CO2 emissions is compatible with the NEM's energy-only market design.

Australia's renewable energy target (i.e., renewable portfolio standard) required that energy retailers meet a 20% renewable market share by 2020. The certificated scheme will be comfortably met—by 2018 the NEM had 8000 MW of new VRE plant (5000 MW wind, 2800 MW solar) and a further 5500 MW was under construction as at the start of 2019 (2500 MW wind and 3000 MW solar). On-market transactions had delivered more than \$20 billion of investment without any need for governmen<sup>t</sup> involvement or intermediation beyond setting the policy framework. The market was given a target, and market participants (i.e., utilities, investors, and customers) delivered the capacity.

Used carefully, CfDs present policymakers with a reliable tool which can be used to overcome an array of market failures, including those associated with missing or incomplete markets (including emergency plant for security of supply reasons, certain positive, or negative externalities including CO2 emissions, R&D and externalities arising from first-of-a-kind commercialization investments). In the NEM, CfDs have been used selectively and e ffectively by state governments to "prime" emerging markets, navigate Commonwealth Government policy discontinuity, with material on-market transactions following. The Australian Capital Territory governmen<sup>t</sup> CfDs pioneered nominal-price transactions, the Queensland government's CfDs led to more than 1900 MW of follow-on solar PV projects, and the SA government's semi-CfD for battery storage in has since resulted in more than a dozen battery projects either under active development or commitment. From a project execution perspective, the e ffectiveness of government-initiated CfDs are unquestionable.

But in an energy-only market setting, government-initiated CfDs must be used judiciously because they introduce "quasi-market participants" who do not respond to spot market signals, and do not participate in forward markets at all. Quasi-market participants are indi fferent or substantially immune from future outcomes in spot and forward markets. This can result in plant entry that is poorly timed, poorly sized, poorly located, and above all, poorly motivated to respond to the electricity and frequency control ancillary service spot price signals which keep the power system operating in a stable manner.

CfD plant also benefits from otherwise unachievable credit metrics owing to a taxpayer-financed and credit-wrapped CfD instrument—with the risks transferred to taxpayers. Prima facie this may deliver lower project specific LCoE's, but this is only because risk has been transferred to taxpayers (which is not costless), and whether this is an e fficient allocation and use of governmen<sup>t</sup> balance sheet capacity is, in my opinion, questionable. A wide-ranging program of government-initiated CfDs can be expected to crowd-out on-market rival merchant/bilateral investments.

If there is an upside to the present analysis, it is that the number of alternate policy instruments available to governmen<sup>t</sup> to achieve policy objectives has expanded very rapidly [53]. A wide array of policy instruments exist to deal with the market failures which CfDs are intended to remedy; renewable energy policy objectives can be achieved by an emissions obligation or well-designed renewables

portfolio standard [48]; the need for emergency capacity can be (and in the NEM recently has been) dealt with by establishing minimum exit notification periods for plant intending to exit the system; resource adequacy (i.e., adequate plant capacity including an appropriate reserve plant margin) can be maintained by ensuring the level of VoLL remains appropriate or by pursuing reliability obligations if this becomes necessary. All of these options work with an energy-only market design, including the forward market for contracts.

Throughout most of 2018, Australian policymakers developed a policy known as the National Energy Guarantee which had two embedded policy mechanisms for energy retailers to comply with; (i) an emissions obligation which was consistent with Australia's international CO2 commitments under the Paris Agreement, and (ii) a reliability obligation which was consistent with the NEM's reliability criteria and was designed to ensure resource adequacy. The former was designed to encourage hedge contract activity with new renewable projects, and the latter was designed to be acquitted via ensuring adequate forward contracts were committed—both mechanisms were thus designed to add to liquidity rather than detract from it.

In contrast, as quantitative results and analysis in this article explain, a wide-ranging program of government-initiated CfDs can be expected to impair market e fficiency in an energy-only market setting. While adding to buy-side liquidity, government-initiated CfDs replace on-market transactions and thus subtract from sell-side liquidity, and this matters in loosely-interconnected energy-only markets because as coal plant exits, *primary issuance* hedge contract shortages become predictable. Shortages in the forward markets may harm consumer welfare by raising contract premiums—the primary input into consumer tari ffs—and by forcing the most price-elastic industrial customers into accepting spot market exposures, which at best disrupts manufacturing e fficiency. Further, CfD-driven hedge market shortages may unintentionally foreclose non-integrated 2nd tier retailers—deeming such a program of government-initiated CfDs to be (unintentionally) anti-competitive. Consequently, the *National Energy Guarantee* or an equivalent suite of policies seems a better place for Australian policymakers to focus on.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The author declares no conflicts of interest.
